Previous volumes inthis series ‘Volume 1 Geometrical theory of diffraction for electromagnetic waves Graeme L James Volume 2. Electromagnetic waves and curved structures Leonard Lewin, David C, Chang and Edward F Kuester Volume Microwave homodyne systems Ray J. King Volume 4 Radio dreetiontinging P.J.D. Gething Volume 5 ELF communications antennas Michae! L. Burrows Volume 6 Waveguide tapers, transitions and couplers F. Sporleder and H. G. Unger Volume 7 Rellector antanna analysis and design PJ. Wood Volume 8. Etfecs ofthe troposphere on radio communications Martin PM. Hall I Volume 9 Schumann resonances in the earth-ionasphere cavity PV. Bliokh, AP Nikolaenko and Y.F. Flippov ‘Volume 10 Aperture antennas and difraction theory EV. dull i Volume 11 Adaptive aray principles ‘I.E Hudson a Volume 12 Microstrip antenna theory and design J.B. James, PS. Hal and C. Wood Volume 12 Energy in electromagnetism | HIG. Booker I ‘Volume 14 Leaky feeders and subsurface radio communications, | P.Delogne Volume 15 The handbook of antenna design volume 1 i Editors: A. W. Rudge, K. Milne, A.D. Olver P. Knight 1 Volume 16 The handbook of antenna design volume 2 Editors: A. W. Rudge, K Milne, A. O. Olver P Knight Volume 17 Suneilance radar performance prediction | P.Rohan Volume 18 Corrugated hors for microwave antennas P.J.B. Clarrcoats and A. 0. OWver Volume 19 Microwave antenna theory and design S.Silver i Volume 20 Advances in radar techniques ioe N.Marcuvitz Peter Peregrinus Lid i (On behalf of The Insiitution of Electrical EngineersPreface to 1986 E Errata THIRTY-FIVE yearshave elapsed sine theft printing of the Waves Handbook. i Page Reads Should read Public domain editions by various publishers have appeared in the interim, but the ; Curent “out of print status ofthe Handbook has been cause of lament by interested 32, 4th line from bottom Eqn(21) Ban) students and microwave researchers. The decision by the Institution of Electrical Ath line after Eqn(62): implies, in general jes not imply Engincers to undertake e new printing remedies this condition and comes as personal th tine after Eqn(62): must consequently be ___can be effected in terms plessreand tibet the effort ofthe insvdulsacknoleed nthe ginal peace efected ona wala bass of vector modes with “The new edition also provides an opportunity to correcta partial ist of typographical an 436, Fig. 10 legend vero exhxr, substantive errors pointed out aver the past years by colleagues. Peete (eee x0 infinite facuvire, 401 Tine 2: etuy) xy) NM) 52, Eqn(97a) Je i. reagan Yom | Eang7e i 5 | 47, line 1: Figs Fig..8b | ‘53, end of first paragraph are not as yet are 63, top figure labelled Hig should be reversed ie. 3rd and 4th columns bod of arrows: poy bod i > < 72, Eqn(31): Vo -V'on Too ton 73, Bqa(32}: v v : r - : 79, Eqn at top figure: 1.873 1.946 I ‘88, 2nd line after Eqn(63): falls normally is incident i 104, line 5: Zope * Zeg, 306 RCZg)= 0 Zo * Ze fr rociprocal ' and Re(Znq) = 0 for nondisipaive structures 108, Eqn(10): Sr = Som (for seclprocal structures) 111, Fig3.3 labels: © ©. 0) © 115, Fig 3.7 legend: Heplane E-plane 126, line 2: the shunt the * shuntxii Page 155, Eqa(88a) 155, Bqa(88b): 178, Eqo(1): 182, Fig.46-5 184, Bqn(2a): 196, line above Eqn): 201, line above Eqn(1) 208) line above Eqn) 208, line 3 in Eqn(1): 216, 7 lines from bottom: 224, line 3-4 of Restrictions: 238, line after Eqn( Ib): line 3 after Eqa( tb): 239, line 8 246, line 4 in Experimental Results: line 7 in Experimental Results: 256, Eqn after Eqn(6by 251, Eqa(7) Ean(8) 266, Eqn(la) + (1b): 266, 3d line after Eqa(2b): 289, Fig.5.22-1 Side View: Transpose Reads EQ) HY) EQ) win ror » et No) where by where t range of validity ‘an incident lowest mode tan? bd z dotted sin? nox Pd ‘Should read FG) Hz) Ex) Hz) ("il a» b x =N.@®) where (y= 1.781) by (y= 1.781) where (= 1.781) 1 range of validity, except for a~b, ‘po incident lowest ‘modes 9 tan? ba z solid Page 339, Fig. 6.1-3 Side View: 339, Numerical Results ine 3 Experimental Results tine 3 340, Fig 6.14 3, Fig6.15 342, Fig 61-6 343, Fig6.-7 34, Fig6.1-8 345, Fig619 346, Fig 61-10 347, Fig6 1-11 348, Fig6.1-12 349, Fig61-13 350, Fig.6.1-14 Reads ‘Should read Ty should be below line + ag b de Blok lo lok lok lo 7'™ Blok | de ‘Should corrections other thaa the above be noticed, the publishers and Series Editors of the IBE’s Electromagnetic Waves Series would be grateful if they could be ‘communicated to them eo thatthe errata sheet can be updated in future editions, thus benefiting the next generation of research workers.Contents FOREWORD. . bocce eee * PREFACE, : . vil Cuan, 1, TRANSMISSION LINBS 1 14, Waveguides as Transmission Lines, 1 12, Field Represeatation in Uniform Waveguides a 13, Uniform Tramamiscion Linos. Impedance Descriptions 7 14. Uniform Tramamission Lines, Scattering Deseriptions 1B 15, Interrlations among Uniform Trensmissionline Descriptions. . 16 148, Uniform Tratstnission Linea with Complex Perametars. ra (2) Waveguides with disipation (6) Waveguides beyond eutof! 17, Field Reproseatstion in Nonuniform Redial Waveguides 18, Field Representation in Nonuniform Spherieal Waveguides Cuar.2, TRANSMISSION-LINE MODES 24, Mode Charecerstos bocce eee 22. Rectangular Waveguides (@) Bmodes © H-modes (©, Modes in » Parallel Plate Guide 29, _Girvular Waveguides (@) Emnodes () Hemodes 24, Coaxial Waveguides . : 2 (@) Bmodes (@) H-modee 25. Buiptical Waveguides bebe e eee eee 26. Space as a Usiform Waveguide 8 (@) Field in fre space (@) Field in the vicinity of gratings 27, Radial Waveguides 0 (@) Cylindrical cross-sections ) Cylindtieal sector eros-sections 28 Spherical Waveguides sine 96 () Fields in fe space (©) Contes! waveguides sae 88 & Cnar, 3, MICROWAVE NETWORKS 101 241, Representatia of Waveguide Diacontinuitica 101 @) Impedance representation (@) Admittanes representation (©) Beattering representationwi CONTENTS 32, Equivalent Cireuita for Wavoguide Diseontinuities 89, Equivalent Representation of Microwave Networks 34, Meaurement of Network Parameters 35, Theoretical Determination of Circuit Parameters Cuar.4, TWO-TERMINAL STRUCTURES. Linns Tsmanariva 18 Guipns Bavoxo Cororr 4:1. Change of Cross Section, H-plane . . . « (a) Symmetrical ease (@) Asymmetrical case 42, Bifurestion of « Restangular Guide, H-plane 43, Coupling of « Coaxial Line to a Circular Guide 44, Rectangular to Circular Change in Gross Section 45. Termination of a Coaxial Lino by a Capacitive Gap Lawes Raptarine 110 SPACE ‘46a, Peralle-plate Guide into Space, H-plane “40, Rectangular Guide into Bounded Space, H-plane . - 47a, Parallekplate Guide Radiating into Space, Hplane. 47H, Rectangular Guide Radiating into Bounded Half Space, B-plane 48, Paralleplate Guide into Space, H-plane ‘£9. ParalleL-plate Guide Redisting into Half Space, H-plane 410, Apertures in Rectangular Guide... (@) Rectangular spartan Circular a 4411, Array of Sosnfnte Planes, Aeplane 412, Radiation from a Gzealar Guide, Bermode. 43. 416, Radiation from a Circular Guide, Hirmode. 415, Coaxial Line with Infinite-centar Conductor 416, Coaxial Line Radiating into Semi-ininite Space Crar.5. POUR-TERMINAL STRUCTURES Smaverunns wins Zeno isos ‘1. Capacitive Obstacles and Windows in Rectangular Guide (@) Window formed by two obstacles (@) Window formed by one obstacle (©) Symmetrical obstacle 52, Inductive Obstacles and Windows in Rectengular Guide () Symmetrial window @) Asymmetzicel window (6) Symmetrical obstacle 8. Caparitive Windows in Coaxial Guide (@) Disk on ianer conductor (©) Disk on outer conductor ‘54, Cisoular and Elliptical Aportures in Rectangular Guide. (@) Centered eirelar aperture @) Small elliptical or cioular aperture 55, Blliptical and Cireular Aportures in Gireular Guide 56. Small Elipticel and Cireular Apertures in Coaxial Guide. 108 47 130 138 168, 168 168 1m lam 176 178 179 179 183 183 184 186, . 187 103 195, 198, 201 23 aT 218 a8 221 29 238 246 CONTENTS 57. Annular Window in Ciroular Guide 58, Annular Obstacles in Cireular Guide. Smucrunee wim Fivrrs Trscexnss 59. Capacitive Obstacles of Finite Thickness (@) Window formed by two obstacles (©) Window formed by one obstacle 510, Inductive Wirdowe of Finite Thickness (@) Symmetries! window @) Asymmetrical window 5:11, Solid Inductive Post in Rectangular Guide () Of-centered post @) Centered post (0) Noneievlar posts 45.12, Dielectric Poss in Reotangular Guide 5:18. Capacitive Poet in Rectangular Guide ‘514, Post of Variatlo Height in Rectangular Guide. ‘5415, Spherical Den: in Rectangular Guide 5:16. Circular Obstecle of Finite Thickness in Rectangular Gui 5417, Resonant Ring in Cireular Guide Gravivos anD Amnars 1x Fran Seact 518, Capacitive Stipe 519, Inductive Stipe 5:0. Capacitive Pots 521, Inductive Pov... 5.92, Armay of Semiinfinite Planes, B-plane 158. Array of SemiinGnite Planes, H-plane Asniaezrarc Sreucrunns; Couruive oF Two Gurpes, 524, Junetion of Two Rectangular Guides, H-plane {e) Syrmetrice! case (©) Asymmetrial case 1525, Bifurcation of a Rectangular Guido, H-plane . 1528, Change in Height of Rectangular Guide (a) Symmetrical case (@), Asymmetrical ease 527. Change in Radius of Coaxial Guide (@) Equal outer radi (@) Bqual inner radi 598, F-plane Corners (@) Right-angle bends (Arbitrary angle bends 529, Heplane Cora. (e) Right-sngle bends (©) Arbitrary angle H-plane comers 5:90, Junction of » Rectangular snd x Radial Gide, B.plane ‘531, Coupling of a Coasial and a Circular Guide 532, Coupling of Rectangular and Circular Guide 558. Aperture Cousling of Two Guides... () Junction of two rectangular guides (@) Junction of two citevlar guides () Tunetion of two conxial guides (@) Junction ofa rectangular and cireular guide aT 29 249 29 . 257 266 268 Lan a 310 an 318 S888uae. 6. Gnas. 7. cman. 8 conreNTs 5:94, Cireuler Bends, E-plane 5:35. Gireuler Bends, H-plane SIX-TERMINAL STRUCTURES 6:1, Open T-junction, plane 62, Slit-coupled T-junction in Rectangular Guide, Blane 63, 120° Y-junetion, plane 64, plane Bifureation 65. Open T-junction, H-plane. 56, Sit-eoupled 7-junetion in Resiangular Guide, 67. 120° ¥-junction, H-plane 68. Aperture-coupled P-junctions, E-plane () Rectangular stub guide @) Cireular stub guide 69, Aperture-coupled T-junction in Rectangular Guide, H-plane (@) Restangular stub guide ) Citoular stub guide 6-10. Aporturo-coupled 7-junotion in Cosxial Guide G11, Bifurcation of « Cossial Line BIGHT-TERMINAL STRUCTURES, 7A, Sit Coupling of Rectangular Guides, B-plane 72, Small-aperture Coupling of Rectangular Guides, E-plane, 7, Aperture Coupling of Coaxial Guides. TA. Sit Coupling of Rectangular Guides, H-plane 7-5, Aperture Coupling of Rectangular Guides, #-plane 78, 0" ¥-junction, B-plane. 77. 0° ¥-junction, H-plano 78. Magio-? (Hybrid) Junction COMPOSITE STRUCTURES Proracantox at Componrsn GorDEs S11, Rectangular Guide with Dislectrie Slabs Parallel to 82, Rectangular Guide with Dielectrio Slabs Perpendicular to E 83, Ciroular Guide with Dielectric Cylindors 84. Coaxial Guide with Dielectric Cylinders 85, Rectangular Guide with “Nonradiating” Sit 86, Rectangular Guides with Ridges 87, Rectangular Guide with Resistive Strip. ‘Taicxnsas Berecrs 88, Capacitive Obstactes of Large Thickness 89, Inductive Obstacles of Lerge Thickness (a) Window formed by two obstacles (@) Window formed by one obstacle 810, Thick Cireular Window S11, plane 7 with Sit Coupling APPENDIX GLossal INDEX. RY. aa 336 387 339 352 353 355 360 302 363 366 388 37a 373 35 arr 378 370 380 383 386 387 387 388 301 395, 206 207 309 402 408 408 an 415, wa 423 CHAPTER 1 ‘TRANSMISSION LINES 1. Waveguides as Transmission Lines—The determination of the cleotromagnetic fields within any region is dependent upon one’s ability to solve explicitly the Maxwell field equations in a coordinate system appropriate to the region, Complete solutions of the field equations, or equivalently of the wave equation, are known for only relatively few ‘types of regions. Such regions may be classified as either uniform or nonuniform. Uniform regions are characterized by the fact that cross sections transverse to a given symmetry, or propagation, direction are almost everywhere identical with one another in both size and shape. ‘Nonuniform regions are likewise characterized by a symmetry, or propa~ gation, direction but the transverse cross sections are similar to rather ‘than identical with one anothe Examples of uniform regions are provided by regions cylindrical about the symmetry dirsction and having planar cross sections with reo- ‘tangular, circular, ete., peripheries. Regions not eylindrical about the symmetry direction and having nonplanar cross sections of cylindrical, spherical, etc., shapes furnish examples of nonuniform regions (ef. Sees. 1-7 and 1'8), "In either case the eross sections may or may not be limited by metallic boundaries. Within such regions the electromagnetic field may be represented as a superposition of an infinite number of standard functions that form a mathematically complete set. ‘These complete sets of functional solutions are classical and have been einployed in the mathematical literature for some time. However, in recent years the extensive use of ultrahigh frequencies has made it desirable to reformulate ‘these mathematical solutions in engineering terms. It is with this reformulation that the present chapter will be concerned. ‘The mathematical representation of the electromagnetic field within a uniform or nonuniform region is in the form of a superposition of an infinite number of modes or wave types. ‘The electric and magnetic field components of each mode are factorable into form functions, depending only on tho cross-sectional coordinates transverse to the direetion of propagation, and into amplitude funetions, depending only on the coordi- nate in the propagation direction. ‘The transverse functional form of each mode is dependent upon the cross-sectional shape of the given rogion and, save for the amplitude factor, is identical at every cross 12 TRANSMISSION LINES [Bro 11 section. As a result the amplitudes of a mode completely characterize ‘the mode at every cross section. The variation of each amplitude along the propagation direction is given implicitly as a solution of a one- dimensional wave or transmission-line equation. According to the mode {in question the wave amplitudes may be either propagating or attenuating along the transmission direction. In many regions of practical imporéanee, as, for example, in wave- guides, the dimensions and field excitation are such that only one mode is capable of propagation. As a result: the clectromagnetio field almost everywhere is characterised completely by the amplitudes of this one dominant wave type. Because of the transmission-line behavior of the modo amplitudes itis suggestive to define the amplitudes that measure the transverse electric and magnetic ficld intensities of this dominant mode as voltage and current, respectively. Tt is thereby implied that, the electromagnetic fields may be described almost everywhere in terms of the voltage and current on an appropriate transmission line. This transmission line completely characterizes the behavior of the dominant, mode everywhere in the waveguide. ‘The knowledge of the real charae- teristic impedance and wave number ofthe transmission ine then permits one to describe rigorously the propagation of this dominant mode in familiar impedance terms. ‘The impedance description may be extended to deseribe the behavior of the nonpropagating or higher modes that are present in the vicinity of cross-sectional discontinuities. Mode voltoges and currents aro Introduced as measures of the amplitudes of tho transverse electric and magnetic field intensities of each of the higher modes. Thus, as before, cach of the higher modes is represented by a transmission line but now the associated characteristic impedance is reactive and the wave number imaginary, ie., attenuating. In this manner the complete deseription of the cleetromagnotie field in a waveguide may be represented in terms of the behavior of the voltages and currents on an infinite number of ‘transmission lines. ‘The quantitative use of such a representation in a given waveguide geometry presupposes the ability to determine explicitly 1. The transverse functional form of each mode in the waveguide cross section, 2. The transmission-line equations for the mode amplitudes together with the values of the mode characteristie impedance and propaga~ tion constant for each made. 3. Expressions for the field components in terms of the amplitudes ‘and functional form of the modes. ‘The above-described impedance or transmission-line reformulation of the Sec, 12] FIBLD RBPRESSNTATION IN UNIFORM WAVEGUIDES — 3 clectromagnetie field will be carried out for a number of practical uniform and nonuniform waveguides. 12, Field Representation in Uniform Waveguides—By far the largest class of waveguide regions is the uniform type represented in Fig. 14. Such regions are eylindrical and have, in general, an arbitrary eross section that is generated by a straight line moving parallel to the sym- metry or transmission direction, the latter being characterized by tbe unit vector z In many practical waveguides the cross sectional geom- etry is described by a coordinate system appropriate to the boundary res ection Lengua vew Fla, '4.—Unform waveguide of abiteary eros eeston. curves although this is not ® necessary requirement. Since the trans- mission-line description of the electromagnetic field within uniform ‘guides is independent of the particular form of coordinate system employed ‘to describe the cross section, no reference to cross-sectional coordinates will be made in this section. Special coordinate systems appropriate to reotangular, circular, and elliptical cross sections, ete., will be considered in Chap. 2.’ To stress th independence of the transmission-line deserip- tion upon the cross-sectional coordinate system an invariant transverse vector formulation of the Maxwell field equations will be employed in the following. ‘This form of she field equations is obtained by elimination of the field components along the transmission, or , direction and ean be written, for the steady state of angular frequency o, as 8B: — page + Avie) Oh x 1), aH, n 1 a Fm inte + fp 9.9 * Oo X Bd Vector notation is employed with the following meanings forthe symbols: E, = Ei(ewra) = the rms electric-iold intensity transverse to the Hawa) H, ‘the rms magnetio-field intensity transverse to the n= Vali propagation corstant in medium =o Ve = 2e/ intrinsic impedance of the medium,4 TRANSMISSION LINES IBe0. 12 eer pester rnevene to as! = ¥ ~ a9 = dja defined such that e+ A= Ase = "The time variation of the field is assumed to be exp (++). ‘The z com- ponents of the electric and magnetic fields follow from the transverse ‘components by the relations ikoB Yer eX 29), ian, ° Ve Ge X Ei). Equations (1) and (2), which are fully equivalent to the Maxwell equations, make evident in transmission-line guise the separate depend- ‘ence of the field on the cross-sectional coordinates and on the longitudinal coordinate 2. The cross-sectional dependence may be integrated out of Eqs. (1) by means of a suitable set of vector orthogonal functions. Functions such that the result of the operation V,¥,- on a function is proportional to the function itself are of the desired type provided they satisfy, in addition, appropriate conditions on the boundary curve or curves ¢ of the erass section. Such vector functions are known to be of two types: the H-mode functions ef defined by ef = — Vay, moa) & where Vib: + kd = 0 abr bomeii wo oy Fe Oo Sif Ry = 05% and the Hemode fundtions defined by el = me XV, hyo’, } (4a) where (4b) \ For « cross-section defined by a reetangular zy coordinate system wend trod ‘whore xp and ye are unit vectors in the # and y ditetions. * The case, = Oris in multiply ponneoted eross sections such as those encour tered in coaxial waveguides, The vanishing of the tangentisl derivative of on ¢ Smplies that % isa constant on each periphery. Sec. 12] FIBLD REPRESENTATION IN UNIFORM WAVEGUIDES 5 where é denotes a double index mn and » is the outward normal to s in ‘the cross-section plane, For the sake of simplicity, the explicit depend- ence of el, ef, %, and % on the cross-sectional coordinates has been omitted in the writing of the equations. ‘The constants Ki; and Kj are defined as the cutoff wave numbers or eigenvalues associated with the guide cross section. Explicit expressions for the mode functions and cutoff wave numbers of several waveguide cross sections are presented in Chap. 2. ‘The functions e¢ possess the vector orthogonality properties [Jcaas- [foraas (ISH) [[eeras-o with the integration extended over the entire guide cross section. ‘The product ee; is a simple scalar product or en Hermitian (ie., complex conjugate) product depending on whether or not the mode vectors are real or complex. "The transverse clectris and magnetic fields ean be expressed in terms of the above-defined orthogonal functions by means of the representation B= Dveel+ Yrroets = Dion + Yvon, and inversely the amplitudes Ys and 1, can bo exprested interme of the re sem [fccas, vee [noes o J [ te meas, n [ [mwas ‘The longitudinal field components then follow from Eqs. (2), (3), (4), and (6a) as (6a) im, = J) Teena, < (6c) jitH. > WOME %. In view of the orthogonality properties (5) and the representation (62), the total average power fow along the guide at z and in the z directions,6 TRANSMISSION LINRS (Bro. 12 nel [ ex Ht mas) = m(, vats Y) via"), where all quantities are rms and the asterisk denotes the complex conjugate. For uniform guides possessing no discontinuities within the guide cross section or on the guide walls the substitution of qs. (6a) transforms Eqs. (1) into an infinite set of equations of the type Wo tty al, (8) dee —jxir, ‘hich define the variation with z of the mode amplitudes Viand I, ‘The superecript distinguishing the mode type has been omitted, since the equations are of the same form for both modes. ‘The parameters x; and Z are however of different form; for B-modes geVPoR mat (0) for H-modes gis Vie= (08) Equations (8) are of standard transmission-line form. ‘They. con- stitute the basis for the definition of the amplitudes V; as mode voltages, of the amplitudes J; as mode currents, and concomitantly of the para~ meters x and Z as the mode propagation constant and mode char- acteristic impedance, respectively. ‘The functional dependence of the parameters x and Z on the cross-sectional dimensions is given in Chap. 2 for several waveguides of practical importance. ‘The field representation given by Eqs. (Ga) and (8) provides a general solution of the field equations that is particularly appropriate for the description of the guide fields in the vieinity of transverse discontinuities —such as apertures in transverse plates of zero thickness, or changes of cross seotion. The field representation given in Eqs. (6a) is likewise applicable to the description of longitudinal discontinuities—such as obstacles of finite thickness or apertures in the guide walls. However, as {a ovident on substitution of Eqs. (Ga) into Eqs. (1), the tranamicsion-line equations (8) for the determination of the voltage and current amplitudes must be modified to take into account. the presence of longitudinal discontinuities within the cross section. This modification results in the addition of z-dependent “generator” voltage and current terms to the right-hand members of Eqs. (8). ‘The determination of the mode See, 13) UNIFORM TRANSMISSION LINES 7 amplitude for the case of longitudinal discontinuities is thus somewhat more complicated than for the case of transverse discontinuities. Both cases, however, constitute more or less conventional transmission-line problems. 1.8. Uniform Transmission Lines.—As shown in Sec. 1-2 the repre- sentation of the electri and magnetic fields within an arbitrary but uniform waveguide (ef. Fig. 1-1) can be reformulated into an engineering deseription in terms of ar infinite number of mode voltages and currents. ‘The variation of each mode voltage and current along the guide axis is, described in terms of the corresponding variation of voltage and current along an appropriate transmission line. ‘The deseription of the entire ficld within the guide is thereby reduced to the description of the eleo- ‘rieal behavior on an infinite set of transmission lines. In this seetion ‘two distinctive ways of describing the electrical behavior on a trans- 1nission line will be sketched: (1) the impedance (admittance) deseription, (2) the seattering (reflection and transmission coefficient) deseription. ‘The transmission-line description of a waveguide mode is based on the fact, noted in the preceding section, that the transverse electric field Ey and transverse magnetic field H, of each mode can be expressed as E(zy.2) = roe | Hey) = Teh), (9) where e(ey) and (zy) are veotor functions indicative of the eross- sectional form of the mode lds, and V(¢) and [(e are voltage and current functions that measure the rms amplitudes of the transverse electric and ragnetie fields at any point z along the direction of propagation, As a consequence of the Maxwell field equations (ef. See. 1-2) the voltage and current are found to obey transmission-line equations of the form aw W 2, Ge = WHEL, a yy a) a7 THY, where, for a medium of uniform dielectric constant: and permeability, k_ fam ex teettmates! fa gEX tr Bas Since the above transmission-line description is applicable to every mode, ‘the sub- and superseripte distinguishing the mode type and number will8 TRANSMISSION LINES [Bro 18 be omitted in this seetion, ‘The parameters &, ke, «, and Z are termed the free-space wave number, the eutoff wave number, the guide wave number, and the characteristic impedance of the mode in question. Instead of the parameters k, kx and « the corresponding wavelengths A, X., and Ay are frequently employed. ‘These are related by 2 2n Fae k ca) ‘The explicit dependence of the mode cutoff wave number k. and mode functions e and hon the cross-sectional geometry of several uniform guides will be given in Chap. 2. Together with the knovledge of the wavelength ) of field excitation, these quantities suffice to determine completely the transmission-line behavior of an individual mode. Since the voltage V and current I are chosen as rms quantities, and since the vector functions e and h are normalized over the cross section in accordance with Eq. (6), the average total mode power flow along the Girection of propagation is Re (VI*). Although the voltage V and current I suffice to characterize the behavior of a mode, it is evident that such a characterization is not unique, Occasionally it is desirable to redefine the relations [Eqs. (10)] between the fields and the voltage and current in order to correspond more closely to customary low-frequency definitions, or to simplify the equivalent circuit description of wavernide discontinuities. ‘These redefinitions introduce changes of the form v ME Vege Taal, (124) ‘where the scale factor 1 is so chosen as to retain the form of the power expression as Re (PI). On substitution of the transformations (12a) {nto Eqs, (11) it is apparent that the transmission-line equations retain the same form in the new voltage V and current. I provided a new charac- teristic impedance 2=IN (128) is introduced. ‘Transformation relations of this kind are generally important only in the case of the dominant mode and even then only when absolute impedance comparisons are necessary. Most trans mission-line properties depend on relative impedances; the latter are unaffected by transformations of the above type. Spe, 13) UNIFORM TRANSMISSION LINES 9 Equations (11) may be schematically represented by the transmission line diagram of Fig. 12 wherein the choice of postive directions for V and Tis indicated, To determine explicit solutions of Eqs. (11) it 3s con ‘venient to eliminate either F or V and thus obtain the one-dimensional wave equations cal ares gt (13a) Fter-o (3b) Equations (13) define waves of two types: either propagating or atten- uating with the distance depending on whether the constant x is either positive or negative. Although both types of waves can betreated —_1 a by the same formalism, the follow- ing applies particularly to tho propagating type. Impedance Descriptions —The —" solutions to Eqs. (18) can be written asa superposition of the z trigonometrical functions * 7% ‘gon Fro, 1:2-—Chaioe of postive directions of os xe, sin xz (14) _yollage and errata tier rvaminon By means of these so-called stand” ing waves, the solutions to Eq. (11) can be expressed in terms of the voltage or current at two different points zo and 21 as Ve) = Hee (150) ep) = Hedin des 2) 4 He) an de = 2) (8) ‘or in terms of the voltage and current at the same point 2» as V(e) = Vee) cos ule — 2s) — j21(ce) sin ale — 2), (160) (2) = (G0) 008 «(e — 20) — JY V (20) sim x(z ~ 2). (168) Equations (16) represent the vollage and current everywhere in terms of the voltage and curreat at a single point zs. Since in many applica- tions the absolute magnitudes of V and J are unimportant, it is desirable to introduce at any poin: z the ratio LI@ PVG) _ 2G) Y@= an10 TRANSMISSION LINES called the relative, or normalised, admittance at z looking in the of increasing z, In terms of this quantity Eqs. (16) can be reexpressed, by division of Eqs. (162) and (16), in the form = it YG) cot alee = 2), cot xe — 2) +) v'@) ag) which is the fundamental transmission-line equation relating the relative admittance at any point 2 to that at any other point 2, Many graphical schemes have been proposed to facilitate eomputa- tions with Eq. (18). One of the more convenient representations, the so-called circle diagram, or Smith chart, is shown in Fig. 13. For real « this diagram ropresonts Eq. (18) as a constant radius rotation of the complex quantity ¥’(z») into the complex quantity ¥'(2), the angle of rotation being 2e(2% — 2) radians. Since graphical uses of this diagram have been treated in sufficient, detail elsewhere in this series,! we shall consider only a few special but important analytical forms of Eq. (18) For P'(%)) = =, ¥"(2) = ~j cot ales — 2); (19a) for ¥"(e) = 0, Y@ = +) tan alee = 2); (196) for ¥"(z9) YG) = 1 (19) ‘These are, respectively, the relative input admittances at # corresponding toa short circuit, an open circuit, and a “match” at the point 2. ‘The fundamental admittance relation (Bq. (18)] ean be rewritten as an impedance relation FO) = cot (20 — 2) + 9Z"(20) 0) ‘The similarity in form of Hgs. (18) and (20) is indicative of the existence of a duality principle for the transmission-line equations (11). Duality in the ease of Eqs. (11) implies that if V, I, Z are replaced respectively by 1, V, Y, the equations remain invariant in form. As a consequence relative admittaneo relations deduced from Eqs. (11) have exactly the same form as relative impedance relations. Tis occasionally desirable to represent the admittance relation (18) Ly means of an equlvalent circuit. The circuit equations for such @ representation are obtained by rewriting Eqs. (16) in the form TC) = i¥ cot aes ~ {V1 ~ 3¥ ese aler ~ 1—VCCD!, | aya) Mee) = ~JY¥ ose x(z0 — 2)[V()] — FY cot ules ~ 2) —VGo)} Gf. G.L. Ragen, Mierowase Transmission Circuits, Vol 9, Redistion Laboratory Series Seo, 13) UNIFORM TRANSMISSION LINES u ‘tenuation in 1 decibel steps Pivot at center of calculator slider for arm Fro, 19.—Ciede diagram for uniform transmission lines ‘The equivalent circuit is schematically represented by the x network shown in Fig. 1-4 which indicates both the positive choice of voltage and current directions as well as the admittance values of the circuit elements for a length I = z¢ ~ 2 of transmission line. By the duality replacements indicated above, Eqs. (21a) may be ‘written in impedance form as V(2) = —J2 cot e(ec — 2)[T(z)] — 7Z ese x(z6 — z)[—I(20)], V(ee) = 32 ese alee — UNG] — 32 cot x(z0 — 2)[—Ten | OH)12 TRANSMISSION LINES (Bro. 18 Hence an alternative equivalent circuit for a length 1 of transmission line may be represented by the T' network shown in Fig. 1-4b wherein are indicated the impedance values of the circuit elements. ‘The relation between the impedances at the points 2 and 2» follows from the above circuit representations by the well-known combinatorial rules for impedanees. 20), re. vey @ o Fie, b4-—(@) wCiteult for a length 1 of uniform tranamiaron ne; (6) Tesceit for » ° Tongs f anor immo ln. An alternative form of Eq. (18) useful for conceptual as well as ‘computational purposes is obtained on the substitution V'@ = ~J oot H(2). (220) ‘The resulting equation for #(2) in terms of 8(ze) is, omitting an additive multiple of 2x, (2) = 6(20) + x(ze — 2). (2b) “Tho quantity 9(2) represents the electrical “length” of short-circuit line equivalent to the relative admittance Y(@). ‘The fundamental transmission-line elation (18), expressed in the simple form of Bq, (226), states that the length equivalent to the input admittance at = is the algebraic sum of the length equivalent to the output admittance at zy plus the electrical length of the transmission line between ¢ and zs. It should be noted that the electrical length corresponding to an arbitrary admittance is in general complex. In addition to the relation between the relative admittances at the two points ¢ and to the relation between the frequency derivatives of the relative admittances is of importance. ‘The latter may be obtained by differentiation of Eqs. (21) either as + (28a) or, since from Bq. (115) «(8-07 Seo, 1] UNIFORM TRANSMISSION LINES 1B ay) _ (RP de dY"(e) Teaver 4 @ MeO o tre pve 2) Tt should be emphasized that Hq. (286) determines the frequency deriva- tive of the relative admittance. If the characteristic admittance ¥ varies with frequency, it is necessary to distinguish between the frequency derivatives of the relative admittance ¥"(2) and the absolute admittance ¥(2) by means of the relation of « [6%] + ro (02) ee de Equations (22) to (24) are of importance in the computation of frequency sensitivity and Q of a waveguide structure. 14. Uniform Transmission Lines. Scattering Deseriptions.—The scattering, just as the impedance, description of a propagating mode is based on Eqs. (10) to (11), wherein the mode fields are represented in terms of a voltage and a current. For the scattering description, ov: ‘ever, solutions to the wave equations (13) are expressed as a superposition of exponential functions om and ete, (25) which represent waves traveling in the direction of inereasing and decreas- ing 2, ‘The resulting traveling-wave solutions ean be represented as Vee) = Vine eH + Veg ted, (26a) ZU(@) = Vin EHO — Vo e103, (260) where Vas and Vue are the complex amplitudes at 2 = zo of “incident” and “reflected” voltage waves, respectively ‘Equations (26) constivute the complete description of the mode fields ‘everywhere in terms of the incident and reflected amplitudes at a single point, Since many of the physical properties of the mode fields depend only on a ratio of incident and reflected wave amplitudes, it is desirable to introduce at any point z the ratio V, Pi called the voltage reflection coefficient. The current reflestion coefficient dofined as the nogative of tho voltage reflection coefficient is also employed in this connection. However, in the following the reflection coefficient is to be understood as the voltage coefficient. In terms of Eqs. (26) and (27) the expression (7) for the total average power flow at any point z on a nondissipative uniform transmission line becomes T@) = @)uM TRANSMISSION LINES (Seo. M4 Wel — ir, aa) Wad Z Zz Z which may immediately be interpreted as the difference between the incident and the reflected power flowing down the guide. Equation (28) makes evident the significance of |I}* as the power reflection coefficient, which, in turn, implies that [T| <1. ‘The relation between the reflection coefficients at z and 2» is simply T@) = Pees, (29) P= Re (VI*) ‘A graphical representation of Eq. (20) is afforded by the circle diagram shown in Fig. 1 from which both the amplitude and phase of the re- flection coefficient’ may be obtained. The greater simplicity of the fundamental refletion-cooffcient relation (20) es compared with the ‘admittance relation (18) implies the advantoge of the former for eomputa- tions on transmission lines without discontinuities. ‘The presence of dis- continuities on the line leads to complications in description that usually ‘are more simply taken into account on an admittance rather than a reflee- ‘tion-coeffcient basis. In any case both methods are equivalent and, as seen by Eqs. (26) and (27), the conneotion between them follows from the relations nyo A= P®) 1 «LY, v@-to oo re ip. (30) It is frequently useful to employ a cirouit representation of the con- neotion betwoen the seattering and impedance deseriptions at any point 0 0f a transmission line. This representation is based on the fact, evie dent from Egs. (26), that VG) = 2Vine — Z1(@9), Gta) (ex) = Bw — ¥V (2), (au) where Tine = ¥Viw ‘These relations are schematically represented by the circuits shown in Fig. 1-5a and b. Figure 15a indicates that the excitation at 2 may be thought of as arising from a generator of constant voltage 2Vine and internal impedance Z. ‘The alternative representation in Fig. 1-56 shows the excitation as a generator of constant current 2/i., and internal admittance Y. A transmission-line description that is particularly desirable from the ‘messurement point of view is based on the standing-wave pattern sel up by the voltage or current distribution along the line. From Eqs. (26a) See. 1-4] UNIFORM TRANSMISSION LINES 15 \de of the voltage pattern at any point 2 is given by Vael VIF TPF BRT cos), (32) re) = [rlee, where defines the amplitude [P| and phase ¢ of the reflection coefficient. Most probe types of standing-weve detectors read directly proportional to the voltage amplitude or its square. ‘The ratio of the maximum to the mini Ke) 10) : J, Ta | t x 2c "God 2 eo @ o io. 16.—(o) Repreentation ofan inoident wave at zy ara constant-voltege generstor; (8) copreseneeion of an Snodont wave ae seas m sonsnircurrent generater, ‘mum voltage amplitude is defined as the standing-wave ratio r and is given by Eq. (81) as 1, L+ T= (33a) and similarly the location of the minimum zu is characterized by Been) = (336) At any point « the relation between the reflection evellicient and the standing-wave parameters can then be expressed as 1 rH rT ya Shem, @ T@) = oie) = For the calculation of frequency sensitivity it is desirable to supple- ‘ment the relation between the reflection coefficients at two points on a transmission line by the co:responding relation for the frequency deriva- tives. The latter is obtained by taking the derivative with respect to « of the logarithm of both sides of Eq. (20). This yields ar(@) _ are) ae TG Tap tate — 2) & 35)16 TRANSMISSION LINES (Sno, 15 For the case of nondissipative transmission lines (« real) it is useful to separate Eq, (35) into its real and imaginary parts as a\rG@s)| are)) (86) (86) since from Eq, (118) de _ (8)*ak _ (BY dew TOM) EM) we It is seen that on a relative change of frequency du/a, the relative change dP\/IC| in amplitude of the reflection coefficient: is identical at any two points 2 and 25 on the transmission line. The absolute change a in phase of the reflection coeflicient at 2 differs from that at ze by an amount proportional to the change in electrical length of the intervening line, Equations (85) and (86) are equivalent to the corresponding Eqs. (23) for the admittance frequency sensitivity. ‘The former are more suited for the investigation of broad-banding questions on long transmission lines, while the latter are more suited to the computation of Q's of short lengths of transmission lines or cavities. 4.6. Interrelations among Uniform Transmission-line Descriptions. ‘The interrelations among, the impedance, relative admittance, refletion coefficient, and standing-wave characterizations of the voltage and ctr rent behavior on uniform transmission line may be summarized as r-i l= w-4 rein =— pai" -iey ger Bia yet li-t + 7 cot xd, Vos = TEP* cota — ir G7) On separation into real and imaginary parts these relations may be writ- ten in the form [a=oy ee" +x" PFI" Vageya a” Vp x = dad +r = tans (28 + Gea) 2x? = tant (— 2X" _), 3) “ te =) s») (Carer Ir) (88a) See. 15] UNIFORM TRANSMISSION LINES Ww pa LeI) VO ROR ER + VOTO a THI” Vasey +B — Va—6) + Bt _ V@ PE + VRE (38) Ve +i eX Va xP 7 Rt _* a — [ne C= pep ge Painted Fema ~ TF Aes eT RY CS) (2 = 1) oot ud =2)n) sin & (ase) “FF cot ad ~ TF OP cos FT : 1 =n? Feed + int ad . =r) cot ad pe Feta ti where Y= G’ + JB’ B= RLGX Taare tar | 8) 2ir| sin ® =O cos & FTF (G80) relative admittance at 2 ‘lative impedance at 2. T= [Plo = reflection coefficient at z d P, P, voltage standing-wave ratio. distance to standing-wave minimum. relative transmitted power. relative reflected power. As previously stated Fig. 13 provides e graphical representation of most of the above relations. In addi- tion the graph of the dependence of P,, Pr, and [P| on +, shown in Fig. 16, is often of use. 16. Uniform Transmission Lines with Complex Parameters. a. Waveguides with Dissipation — ‘The presence of dissipstion in cither the dielectric medium or metallic walls of 2 waveguide modifies slightly the transmission- line deseription (Eq. (11)] of a propagating mode. This modifi- cation takes the form of a com plex rather than an imaginary 19 as} oa| as} Sos = oat oa} 2| | dl Vi ie 2s ate vow Fro, 15.—Relation between VSWR and (oy retiction convient 1, () melative power Feteted Ps, (0) relative power tanamited Pe Propagation constant y and leads to transmission-line equations that may be written (38)18 TRANSMISSION LINES ino. 16 ‘The complex propagation constant y may be expressed as yeatii~ VERB, (390) where the attenuation constant a, the inverse of which is the distance along 2 for the field to decay by 1/e, and the wave number 6 = 2/4, are determined by the type of dissipation, the mode in question, and the geometry of the waveguide. The quantity 8.686a, the decibels of attenuation per unit length, or its inverse 1/8.686a, the loss length per decibel of attenuation, is frequently employed as a measure of attenuation instead of a. ‘The characteristic impedance Z = 1/Y is likewise complex and, for the same voltage-current definitions (10) as employed in the nondissipative case, is given by TYE for H-modes Mm (390) = % for B-modes, Gwe where and «, the permeability and dielectric constant of the medium filling the waveguide, may in general be complex. lectric-type dissipation in the dielectric medium of a waveguide may be taken into account by introduction of e complex relative dielectric constant. ~ ie, (40) where «is the relative diclectrie constant and ¢” the loss factor. For a ‘medium having a relative permeability of unity, the propagation constant, 1-G)-®s an In a waveguide having « cutoff wavelength %, > de the attenuation constant « is, therefore, ea! oe (ETE OE AM oe oO Oe z Hoo 2 - 2 1, and are expressed in terms of the skin depth 3 rather than ¢”. In each case the leading term provides a good approximation for most of the dielectrics and metals encountered in practice.22 TRANSMISSION LINBS (Bre. 16 Measurements of the loss factor ¢” and conductivity ¢ at various wavelengths are displayed in Tables 1-1 and 1-2 for a number of dielectrics ‘and metals. ‘The conductivity properties of a nonmagnetic metal are Fro, 17-—Phase and attenuation fonotions ya = frequently described by its characteristic resistance &, which is related to its skin depth 6 and conductivity © by ane yf s = soa 10 (CE) E cams « being measurod in bos per mete, and sin mete. To facilitate computations of a and @ a graph of the functions woh (282) is (38) ust 1 23 My jy ow ie a“) may be introduced to account for dissipation of « magnetic type. For medium of unit relative dielectric constant, the attenuation constant and wave number may be obtained from Eqs. (42) and (48) by the replace- ment of ¢' and ¢” by »’ and ", respectively. ‘The skin depth 6 in this caso is (45) ‘where the conductivity #« aecounts for the magnetic type of dissipation, Extensive tables of the loss factor 4” or alternatively the conductivity am are not as yet available, ‘The presence of dissipation of both the electric and magnetic type may be taken into account by-introduction of both a complex relative dielectric constant and a complex permeability, as given by Eqs. (40) and (44). The attenuation canstant and wave number may again be obtained from Eqs. (42) and (43) if¢ and «” therein are replaced by eu’ ~ e's” and e's! + ea", respectively. In this ease the skin depth 3 is given by (40) ey a om oe ‘When the medium is an it complex conductivity where od gas, it may be desirable to introduce a a= ~ ja” 7) to describe both the dielectrie and dissipative properties of the medium. For a medium of unit relative dielectric constant and permeability the attenuation constant and wave number can be obtained from Eqs. (42) and (43) on the replacement of ¢“and e” therein by 1 —c"/ote and ¢ fae, respectively. ‘The characteristic admittance of @ propagating mode in » dissipative guide follows from the knowledge of the complex propagation constant. For example, in a dissipative dielectric medium the characteristic admit- tance for H-modes is given by Eqs. (396), (42), and (43) as [oo (S282) — jim (32424 TRANSMISSION LINES [See. 1-6 ¥ fae [a -s(0e fey, (480) i ho Xe o 3, Eq. (51) states in general that Y"(@) 21 independently of the value of ¥"). Although Eq. (61) provides a straightforward means for admittance computations in dissipative transmission lines, such computations are tedious because of the complex nature of the propagation constant, In ‘many practical problems dissipative effects are slight and hence have a small, albeit important, effect on admittance calculations. For such problems a perturbation method of calculation is indicated. In this ‘method one performs an admittance calculation by first assuming the propagation constant to be purely imaginary, ic, = JB as for the case of no dissipation; one then accounts for the presence of dissipation by aulding the admittance sorrection rue tn a pertuirhation ain 7. ‘Thue in the case illustrated in Hg, (623) one notes that the input admittance of a short-circuited length of slightly dissipative line is the sum of the unper- turbed admittance Yo = coth jét and the correction (dYo/dy) «due to the perturbation arin 7. Equivalent-cireuit representations of Eq. (51) can be obtained from See. 146] UNIFORM TRANSMISSION LINES 2 ‘those in See. 1°8 (cf. Fig. 1-4) by the replacement of « therein by ~jy. “Another useful representation of the equivalent network between the input, ‘and output points of a dissipative line of length J consists of a tandem connection of a nondissipative line of electrical length pl and a beyond ‘cutoff line of electrical length —jal, the characteristic impedances of ‘both lines being the same as that of the dissipative line. ‘The scattering description of a propagating mode in a dissipative guide is based on wave functions of the type em and ote, ‘These functions represent waves traveling in the direction of increasing and decreasing 2, respectively, and attenuating as e*™, A mode description can therefore be expressed in terms of an incident and reflected wave whose voltage amplitudes Vix and Vx are defined as in Eqs. (26) with « replaced by ~jr. A reflection coefficient, Te = ye, (3) may likewise be defined such that at any tivo points z and z» TE) = Peden (oH) However, the total power flow at 2 is now given by (65) P = ReiVI*) (: = ine = 2 where Pow = Yo|Vielt 0, ‘The subscripts r and i denote the real and imaginary parts of a quantity, and ¥ is the complex characteristic admittance for the mode in question. From Eq, (658) it is evident that for dissipative lines [F| ean no longer bbe regarded as the pover-eflection coefficient, Moreover, [P| is not restricted to values equal or less than unity. ‘The meaning of I as a reflection coefficient can be retained if the voltage and current on the dissipative line are defined so as to make the characteristic admittance real; in this event Eq. (155) reduces to the nondissipative result given in Eq, (28). 'b. Waneguides beyond Cutog —The voltage and current amplitudes of fa higher, or nonpropagating, mode in a waveguide ure described by the transmission-line equations (89). In the absence of dissipation the propagation constant is real and equal to A> de (56)28 TRANSMISSION LINES (Seo. 16 ‘The nondissipative decay of the mode fields in decibels per unit length (ame unit as for A.) is therefore ust fr RY, a> 6 At low frequencies the rate of decay is independent of d, the wavelength, of field excitation, and dependent only on the geometry of the guide ‘ross section. Values of the cutoff wavelength A, are given in Chap. 2 for several waveguide modes and geometries ‘The characteristic impedance of a beyond-cutoff mode (ie. 2 > \) may be obtained from Eqs. (398) and (56) as £ bog : Vv ai Zz for H-modes, (68) WO and is inductive for H-modes, capacitive for H-modes. The knowledge of the propagation constant and characteristic impedance of a beyond-cutoff mode permits the application of the trans- mission-line analysis developed in Secs. 1-3 and 1-4, provided « therein is replaced by jy (y real). The impedance description is given by Ea, (61), and the seattering description by Hig. (54). Several modifications resulting from the fact that + is real and Z is imaginary have already bbeen discussed in Sec. ‘The presence of dissipation within the dielectric medium or the walls of a beyond-cutot waveguide introduces an imaginary part into the propagation constant y. If dissipation is present only in the medium and is characterized by a complex dielectric constant, as in Eg. (40), we have for the propagation constant = a + J for B-modes, Y (60a) (590) and (60) 608) See. 17] NONUNIFORM RADIAL WAVEGUIDES 29 where Re ‘The approximations (599) and (605) apply to the case of small dissipation with ¢’/¢ <1 and \ not too elose to \.. Equations (59) and (60) for a beyond-cutoff mode and Eqs. (42) and (43) for a propagating mode differ mainly in the replacement of the attenuation constant of the one ease by: the wave number of the other ease and conversely. ‘This correspondence 2p 6 r 3] ‘Side view Side view ‘Top view ‘Top view (adoyinsiat @) Sectoral, Fra, £8.—Radial waveruides between the two cases is general and applies as well to the other types of dissipation mentioned in Sec. 16a. 1:7, Field Representation in Nonuniform Radial Waveguides.—Non- uniform regions are characterized by the fact that cross sections trans verse to the transmission direction are similar to but not identical with one another. A radial waveguide is a nonuniform cylindrical region deseribed by an rée coordinate system; the transmission direction is long the radius r, anc the eross sections transverse thereto are the #z ypical examples of radial eqlindrical surfaces for whieh r is constant. waveguides are the cylindrical and eylindrical sector regions shown in Figs, 1-82 and b. In the rz polar ccordinate system appropriate to the radial wave- guides of Figs. 1-8, the field equations for the electric and magnetic field ‘components transverse to the radial direction r may be written as30 TRANSMISSION LINES (Sse, 17 1 (Lau, _ aH me [ma +h 73g 2 ae), 1/1 oH, 1 eH, (612) wee (LPH) ~- [eC | 8 ty = in| mes A(R 22H] — putt, = 1B: _ OB (02) Te a Be ‘A component form of the field equations is employed because the left-hand members of Eqs. ((1) eannot be written in invariant vector form. ‘The inability to oblain a transverse vector formulation, as in Eqs. (1), implies, in general, the nonexistence of a field representation in terms of transverse aeclor modes. ‘The transverse field representation in a radial waveguide must consequently be effected on a scalar basis For the case vhere the magnetic field has no s-component, the transverse field may be represented as a superposition of a set of E-type modes. ‘The transverse functional behavior of an H-type mode (ef. Sec. 2-7) is of the form sin ain * © where the mode indices m and n are determined by the angular aperture and height of the cylindrical gz cross section of the radial guide. ‘The amplitudes of the ‘ransverse electric and magnetic fields of an E-type mode are character zed by a mode voltage V{ and a mode current I, For the case of ro 2 component of electric field, the fields ean be repre- sented in terms of aset of H-type modes whose transverse form, as shown, in Sec. 27, is likewise characterized by functions of the form (63). ‘The voltage and current amplitudes of the transverse electric and magnetic ficld intensity of an H-typo mode are designated as V1! and I!’. For the ease of a general field both mode types are required, and these are not independent of one another. Incidentally, it is to be emphasized that the above classification into mode types is not based on the trans- See. 1.7] NONUNIFORM RADIAT WAVEGUIDES 31 mission direction. Relative to the x direction all modes are generally hybrid in that they possess both an E, and an Hf, component (¢f. See. 2-7) ‘On substitution of the known transverse functional form of the modes, into Eqs. (61) there are obtained the transmission-line equations (64a) sar for the determination of each of the mode amplitudes V and Z. Because of the identity in form of the equations for all modes, the distinguishing sub- and superscripts have been omitted. The characteristic impedance Z and mode constant x are given by for the E-type modes, (ou) for the H-type modes, fs ~ @) , where 1” and 1" aro constants dependent on the cross-sectional dimen- sions of the radial waveguide and the definitions of V and I (cf. Sec. 2-7). ‘Because of the indicated variability with r of the propagation con stant and characteristic impedance, Eqs. (64a) are called radial transmis- Sion-line equations. Correspondingly the mode amplitudes V and 1 are defined as the rms mode voltage and current; they furnish the basis for the reformulation of the fld description in impedance terms. The vari- ability with r of the line parameters implies corresponding variability in the spatial periodicity of the fields along the transmission direction. The concept of wavelength on a radial line thus loses its customary significance. ‘Impedance Description of Dominant E-type Mode—In practice, the frequency and excitation of the radial waveguide illustrated in Fig.18a are often such that, almost everywhere, only the dominant E-type mode vvith m = Oand n= 0 is prevent, The field configuration of this trans- verse electromagnetic mode is angularly symmetric with E parallel to the saris and Hin the form of circles about the z-axis. ‘The transverse mode fields are represented as Bard) = YO 1@), et (65) Hilrd,2) =32 TRANSMISSION LINES (See. 17 where zp and @» are unit Veetors in the positive = and ¢ directions. ‘The ‘mode voltage V and eurrent I obey Is. (64a) with « = kand Z = tb/2er (¢j.Sec.27). On elimination of Ifrom Eqs. (64a) the wave equation for V becomes 14 (,a¥) ope tals) +" ‘The two independent, standing-wave, solutions of this equation are the Bessel functions (66) Soir) — and No(kr), wherein it is to be emphasized that \ = 2x/k does not in general imply the existence of a fixed wavelength along the direction of propagation. ‘The impedance description of the B-type radial line is based on the above standing-wave solutions; the voltages and currents at the points r ‘and ro follow from Egs. (64a, b) as V(r) = Vireo) Onley) — J2olre) an(z9), 7 ZI (0) ~ Zeller) estan) — iV) Salen), where Coq) = LOM So(ey) = Lwdaitey — Ni), wie), enfey) = 22DNal2) = Nv), 7a ce and Z = {b/2sr an Ze = $6/2sre are the characteristic impedances tn We ft r and ro, rexpectively. These E+ ottagesurent relations maybe 7 7 z schematically represented by the ‘ radial transmission-tine diagram —z of Fig. 19, whieh algo shows the | postive directions of ¥ and Z. Equations (67) may be con- verted to a more convenient frm ie, 18-—Choke of yntiveietin of by introduction of the relative, or pote sod caont edi tasamiton normalized, edmittances _ 21) YO) Balle) _ Yr) FOF #2 8) ro) and Y"(r) Seo, £7] NONUNIFORM RADIAL WAVEGUIDES 33 at the radii r and ro; these admittances are positive in the direction of inereasing radius. By division of Bq. (67) one obtains the fundamental radial transmission-line relation for the lowest E-type mode as nop) dct Yoo te) Uz), MO ~ Cite) + IV CoRGH) (69) where cto) = - culagy = TDNuG) = Niwa) Cte) FOND = ow) nay) ~ Suzy)" (7) Tolz)Noly) ‘o( x) o(y) — 80 few) - TMG) — Maes) ~§%) ~ saey) and aek, y=hr. ‘The ct and Ct functions are called the small and large radial cotangent funetions; their inverses tn and Th are the small and large radial tangent functions, The radial functions are asymmetric. ‘The nature of the asymmetry is evident in the relation ot(z,y) Hey) = —Ct(y,2), my which may be employed to obtain alternative forms of Eq. (68). ‘The radial functions are plotted vs. y — = with y/x as a parameter in the graphs of Figs. 110 to 112, ‘The curves of Figs. 1-10a and Illa apply when y is less than x, whereas those of Figs. 1-10) and 1-116 are for y greater than z. ‘The symmetry of the functions [(z,v) permits the ‘use of the single graph of Fig. 1-12, for both ranges of y/z. Tn addition to the graphs numerical values of the radial functions are given in Tables 1.3 for several values of y/z. These tables are incomplete, as many of the data from which the curves were plotted are not in a form convenient for tabulation. ‘The parametric values y/x ~ 1, but y — 2 Gnite, correspond to the ease of large radii, In this range et(z,y) & Ct(2y) & cot(y ~ z), and Gy) 21. Thus at large radii the radial and uniform transmission- ‘The transmis- sion equations (69) permit the determination of the relative admittance at the input of a line of electrical length y — 2 from a knowledge of the relative admittance at the output. A few examples will serve to illus trate both the use of Eq. (69) and the physical significance of the radial cotangent functions.35 (e<) sorooay OTP aun ag oy ppes a6 wn jo SUMMER payee 40h mg o» @ wee veri=s sour | 9 NONUNIFORM RADIAL WAVEGUIDES ° Hf) (espe z G 9 5 Fi 3 “te > wonemaan 22M ayer ya ou pes oF om yo SAERPAE nd a 2 Prenies 9 ‘ 8 z- g 3 z 3 jo = (emp 2 z s 3437 es quay seat Ot “(e< f) woreanny SEYPMAM ose ya our wipes ed we (<9 worousaies Sooepsapy OE HME OTL TIPDS o_o zev= o- woun NONUNIFORM RADIAL WAVEGUIDES S00. 17] ; (> 0 vormin SNPS oes im su ps at we SEERA! san samk—eT1 8 a 1_z ey sora=$ co SEESE LED + a | | 5 5 7 5 | : wna 2 2 en: 3638 TRANSMISSION LINES (See, 17 See. 17] NONUNIFORM RADIAL WAYRQUIDBS 39 ‘Tapux 10.—Vawues oF mm Raprat, Funorions. IAN) = Nalehdaty) Hew) ~ TING) = NE) wel g 3 ‘ s * s 10 » i | HER) G5) S58] S58 Sy : or ©. 0.6060) " = a te oa eau SO : ie one 2 a1 es, s | 23 0.4007) st u is own 8 s] Q a oe i +t a 0.9900 | 0.9745] 0. n858 0.8730 | 0.690840 TRANSMISSION LINES Teme Lab Sale Waly) ~ Nala Fev) BG) = T@)Ne) = Noo) (Seo. 17 oe 088 See, 17) NONUNIFORM RADIAL WAVEGUIDES a Tanue 138. iia ‘The relative input admittance of an B-type radial line of electrical length y — 2 with a short circuit [Y"(re) = =] at its output end is, ¥'@) = J etl). «7 For an open circuit (¥’(rs) = 0] at the output the relative input admit tance is Yo es} Tn (2,9). (73)42 TRANSMISSION LINE’ [See 17 For an infinitely long E-type line the relative input admittance at any point r, looking in the direction of increasing radius, is found from Eq, (69) by placing ¥/(re) = Land ry = © to be 7 HP (hr), vo HER, os totes : ite 2 etre, where y = 1.781. ‘The input admittance of an infinite radial line is not in general equal to the characteristic admittance! For E-type lines it is seen to be complex with a negative imaginary, i.e, inductive, part. The relative input, admittance looking in the direction of increasing radius for all noninfinite Y"(0), Sulbr) YO = “iy ¢By (78) riys—tun(2) tet YO2-55 for kr & 1. ‘The relative input admittanco is negative imaginary and henee eapacita- tive; it is to be remembered that the output is at a smaller radius than the input and hence the input admittance is counted negatively. Yy Zu-Za 2am Big % z : % F % (@) o Fro, 113.—(e) s-Cirenit eprosantation ofan B-type radial line of electrical length y — 2 (@) Foire reprenntstion af om Hetypa radial ine of leteeal length = 2 An alternative method of determining the relation between the input and output admittances of a radial line is afforded by the equivalent ci ‘cuit representation for a length y ~ = of line. From Eq, (67) one finds that the parameters of the x circuit representation (ef. Fig. 1-13a) of the Sue. 17] NONUNIFORM RADIAL WAVEGUIDES 43 1- Yn = iY [etcom + vt esttew) Ya Ya -i¥s[- tte + af coven} Yu = i VY¥o oats), (78) where 1+ ebay) Ctle,n) Yeu) ‘The est function may be termed the radial cosecant function since in the limit of large 2 and y it becomes identical with the ordinary cosecant funetion. A useful formula for perturbation and frequeney-sensitivity calcula tions, particularly in resonant radial structures, ean be obtained from the differential form of Hq. (69). The differential change d¥’(r) in relative input admittance arising from either a change d¥"(re) of relative output admittance or a relative change di/k in frequency, or both, is given by ( ayn) fe + orl #) a est*(zu) = I+ GOR TOY OR & ~ (aro) i Vitro) | ak) - ( saPor t bv + ce Ghap) i) at (77) where 2 1+KrOF a) = wer Te) — oO For large kr and krs the ratio a(r2)/a(r) approaches unity, and hence in this range fq, (77) and the corresponding viformstine equation (23) become asymptotically identical. With a short circuit at r the right- hand member of Eq. (77) simplifies to vie + jut) 2 [er 09 +50 #] Sheas Scattering Description of Dominant E-type Mode—he scattering description of the dominant E-type mode in the radial guide of Fig. 18a is boeed on the traveling-wave solutions HP(kr) and HPCE) of the radial-wave equation (66). ‘The Hankel function solutions #9 and Hi represent waves traveling in the direction of increasing and decreasing radius and are the analogues of the exponential functions encountered in uniform lines. Tn terms of these functions the solutions of the radial transmission-line equations (64) for the dominant mode volt~ tage and eurreat can be written a8 V (0) = Vow P(r) + Pat) i) GEN) = Vand) + Vall (kr), 7s)“a TRANSMISSION LINES (See. 17 where Vie and Vig are the complex amplitudes of the incident and reflected waves travding in the direction of increasing and decreasing r, rrportrely "the walgy ih the unifrn tne Bay) Troms terete by te of th oa H(z) = (— jr hae, HP) = (hae ™ (80) bee VIEBTENE, oe = ME bea NS) ‘The amplitude ha and phase 7 of the Hankel functions are plotted and tabulated as functions of = for m=O and 1 in Fig. 1-14 and in Table 14. ‘A convenient measure of the voltage and current at any point + is obtained on introduction of a reflection coefficient defined as the ratio of the amplitudes of the reflected and incident waves. ‘There exist two types of reflection coefficients: a voltage reflection coefficient, Vina HPC) _ Vo Flo. 114—Amplitude snd phase of Hankel ‘onetions of order mea a one TH) = 72 PUR) eon (81a) and a current reflection coefficient TQ) = (818) In contradistinetion to the caso of a uniform line the voltage and current, reflection coefficients are not negatives of each other. ‘The fundamental radial transmission-line equations for the reflection coefficients at any two points r and rs of an B-type line follow from Eqs. (81) either as Ty(p) = Py(ra)eftioiter—mee (622) ne Pifrejettner veer (828) By division of Eqs. (79) the relation between the admittance and reflec- tion coefficient at any point r is seen to be ye 1-22r%y 1g) «LED ia YO - Tp Ter (332) Bre. £7] Bee oC SOR ER ose SS oR See NONUNIFORM RADIAL WAVEGUIDES 45 Tame 1 ate) | C2) 6.459 3.325 2.208 1.792 aaa 1.383 1161 ‘or conversely where and YO Zo) = 1 TH) = AOR ne) HOG) yh Y= SY RG ~ Vie 1= YA) Tier 6 1 107) "Ze" Vey46 TRANSMISSION LINES (Sno. 17 With the aid of Eqs. (62) and (83) radial-tine calculations may be per- formed with either the voltage or current reflection coefficient in s manner similar to uniform line calculations. In fact for large kr it is apparent that the two types of calculations are identical, since Ty = —T, and Y,=Y_=¥. ‘The computation of frequency sensitivity on a long radial line is facilitated by a knowledge of the differential forms of Eqs. (82). From Eqs. (82a) and (80) it follows that atv(r) dtreo) , Af 1 1d dk Tele) ~ Tire) +l meal (4) gives the relation between the changes dl'y(r) and dP'y(rs) in the input and ‘output reflection coefficients due to the relative frequency shift dk/k. As for the case of the corresponding uniform line relation (35), Eq. (84) ean be decomposed into an amplitude and phase part from which it is apparent that the rlative change in reflection coefficient is the same at all points of a. nond ssipative line ‘The description of the dominant E-type mode in the waveguide shown in Fig. 182 is somewhat different from that just described. Because of the vankhhing tangential electric field at the guide walls the dominant mode is no longer angularly symmetric and hence m + 0. ‘The transmission-line description is based on the mth-order Bessel and Hankel functions but otherwise is formally identical with that just described. Since no table or plots of the mth-order radial cotangent, ete., functions are available,* no details of this dominant B-type descrip tion will be presented. ‘The transmissionsine deseription of the higher angular modes in cadial lines likewise depends on mth-order Bessel functions, Description of Dominant H-type Mode—Frequently in the regions shown in Fig. 18 the frequency and excitation are such that, almost everywhere, only the lowest I-type mode is present. ‘The field con- figuration of this mode is angularly symmetrie about the z-axis, and hence m = 0. For regions of infinite height in the 2 direction the mag- netic field is parallel to the axis, the electric field lines are circles or circular ares sbout the z-axis, and » = 0, For the particular case of the waveguide shown in Fig. 1-8a the ficld configurations of the dominant B- and H-type modes are, therefore, dual to each other. For regions of finite height there is an additional radial component in the magnetic field of the dominant Hype mode, and n 70. Duality between the fields of the tivo dominan!-mode types no longer exists. However, whatever the height, the transmission-line description of the dominant H-type Gf. HS. Bennett, “Transmission Line Characteristics of the Sectoral Hora, Pree. ERLE, 8,788 (1349). Seo. 18) NONUNIFORM SPHERICAL WAVEGUIDES aT mode in the waveguides of Fig. 1'8 is dual to that of the dominant E-type mode in Fig, 18a. Duality in this case implies that in Pgs. (64a) the V, ZL, and k of the dominant B-type description are replaced by the I, YV, and of the dominant H-type description. ‘The dependence of the characteristic admittance Y and mode constant x of an H-type mode on the cross-sectional dimensions of a radial waveguide are given in See. 27. As a consequence of duality all relative admittance relations derived above for the dominant B-type mode in the guide of Fig. 1-8a are identical with the relative impedance relations for a dominant H-type mode, pro- ied & is everywhere replaced by «. For example, the fundamental input-output impedance relation for a length y ~ z of the dominant -H-type radial line is obtained from Eq. (69) by duality as Gt 2 (rot (zy) ct(zy) Cen) + 52 lew) where z = er and y = «7s, In addition the relative admittance param- eters (76) of the w circuit representation (cf. Fig. 1-13a) for a length yx of a dominant B-type radial line become the relative impedance parameters for the 7’ circuit, representation (ef. Fig. 1-130) of a length y ~ rof the dominant H-type ine. Its also evident that the scattering description of the dominant H-type mode follows from that of the dominant E-type mode by the aforementioned duality replacements. In employing duality one should remember that an infinite admittance, or short circuit, becomes on the duality replacement an infinite impedance, ‘or open eireuit, and conversely. 418. Field Representation in Nonuniform Spherical Waveguides— Another type of nonuniform region that permits a feld representation in ‘terms of an infinite set of known transmission modes is the spherical waveguide depicted in Fig. 1-152 ot b. ‘The transmission direction is along the radius, and the 86 cross sections transverse thereto are either spherical surfaces as in Fig. 1-15a or spherical sectors bounded by cones of aperture 264, and 2% as in Fig. 1-155. The Maxwell equations for the electric and magnetic fields transverse to the radial direction, which is characterized by the unit vestor r5, may be writlen in invariant vector notation as Zo) (85) 12 py = — jy ¢ + =) “GX 1), i a (86a) 1d oa =~ (4 ). (XB), the longitudinal components being expressed in terms of the transverse components as Sh, = Ver (He X to), } SH, = Ver (to X Ei). (a)48 ‘TRANSMISSION LINES [Seo 18 site vow ‘sicbon BA @ ® Fio, 1:16.—Spheril waveguides: (a) Spherical; (2) Conia! ‘The notation is the same as that employed in Eqs. (1) and (2) save that WW and ¥,, the gradients transverse to rp, are defined by Wevind 18, Wey agrne Alternatively, in the spheriesl coordinate system, 7, 6, appropriate to these geometries, the field equations for the transverse components may be written as a ; fa ain 12 py = ~ate {te + ps3 aha & one Ho 2 36 1 2.2 po = ~ate [10+ pha [ata aa 00 Mo 7) See. 18) NONUNIFORM SPHERICAL WAVEGUIDES 49 and for the longitudinal components as rao [ine ty - 3 rs [3% — J cone 25} ‘The dependence of the fields on tho transverse coordinates may be integrated out because the knowledge of the boundary conditions on the curve or curves s (if any) bounding the transverse cross sections implies the knowledge of the form of the transverse mode fields, As in the case of a uniform region this may be done by introduction of an infinite (68) set of orthogonal vector functions which are of two types: e/(@,¢) and (0,8). ‘The H-mode functions e/(0,4) are defined by =r Vb, Rim te X eb } (80a) where PV Wi + Kobe = 0, %=0ons it 0, Booms itty =o, em) ‘The H-mode functions e/(6,9) are defined by ef er x 9%, winkne| 900) where PV PU +R = 0, oy, HO ons, (008 The two-dimensional scalar operator r°V, -.¥ is represented in spherical coordinates by 2 ino 2 a PVA mG S80, + slag oe ‘The subscript i denotes the double index mn and is indicative of the two-dimensional nature of the mode funetions. The vector v is the outward-directed normal to s in the plane of the cross section. For unbounded eross sections, as in Fig. 1-16a, the boundary conditions on & and ¥; are replaced by periodicity requirements, ‘The explicit dependence of the mode functions e{ and e/ on the cross50 ‘TRANSMISSION LINES (Sno. 18 sectional coordinates @ ¢ will be given in See. 28. At this point we shall state only that the e functions possess the same vector orthogonality properties as in Eq, (€) for the uniform case. In the spherical ease the domain of integration is the entire spherical transverse cross section having the angularsur‘ace clement dS = sin @d0d@. ‘The representation of the transverse field in terms of the above set of vector modes is given by nono Senna Smet Hino = Dono) M+ Yea and that of the longitudinal ficlds follows from Tags. (8604 to (91) as (1) (92) ‘The mode amplitudes Vs and [; are obtained from the orthogonality properties of the mode functions as ve [fmceas (@3) hk J [emas, -vhere since the amplitude relations apply to both mode types, the mode superscript is omitted "The substitution of Bigs. (91) into (86a) and tse of Eas. (6) lead to the defining equations (omitting the mode sub- and superscripts) av int, a 08) a__, a iV, ‘which determine the variation with r of the as yet unknown amplitudes V and I. As before, these equations are of transmission-line form and constitute the basis fer the designation of Y and I as the mode voltage ‘and current. ‘The propagation wave number « and characteristic imped- ance Z of the ith mode are given by Seo. 18) NONUNIFORM SPHBRICAL WAVEGUIDES 5 [Bro tt = min +3) for H-modes, (05) for B-modes, where the numerical value of n is determined by the crose-ceotional shape and the # dependence of the mode in question, Spherical Transmission Lines—The frequency and excitation of @ spherical waveguide of the type illustrated in Fig. 1-15 may besuch that the fields therein are almost everywhere characterized by only a single mode, For such » mode the clestric and magnetic fields transverse to the transmission direction r may be represented as Brad) = Vie) £29), HGrt,8) = 10) BM, where the mode voltage V and current I obey the spherical transmission- line equations (94), and where e and h are known orthogonal vector functions characteristic of the cross-sectional form of the mode. The knowledge of the latter functions reduces the problem of field description to that of the determination of the behavior of V and J on a spherical transmission line. Spherical transmission lines are distinguished by the numerical value of n {of. Eqs. (95)), this mode index being indicative of the 9 variation of the mode fields. In the waveguide of Fig. 1-15a the dominant mode is a dipole field characterized by n = 1, whereas in the waveguide of Fig. 1:15) the field of the dominant mode is angularly sym- metric and n = 0 (¢f. Seo. 28). Although a transmissioncline analysis of these dominant modes can be presented along the lines developed in the preceding sections, a detailed treatment will not be carried out because propriate numerical tables. ‘The transmission-line behavior of typical spherieal modes will, however, be sketched. ‘The wave equations that describe the mode behavior on a spherical ‘transmission line may be obtained for the case of H-modes by elimination of Vi from Eqs. (94) as er. fas 0, [emer ) and forthe H-modes by elimination of 1s on i (960)22 TRANSMISSION LINES (See. 18 ‘The corresponding equations for Vj and 1,’ are not so simple, and hence these amplitudes are best obtained from I and V/' with the aid of Eqs. (G4). Solutions to Eqs. (96) may be expressed in terms of the standing JInlkt) and BC), (970) (Hey (33) ose Gay ‘The functions J,(z) and H(z) are closely related to the half-order Bessel functions; typical functions are Bila) = — 008 2, cose +88, MG) ~ — sine 282 FO here BAG) Toa) = sin x, Jie In terms of theso solutions the current and voltage of an H-mnode at any point r of « spherical line follow from Eqs. (94) as I) = Teoa)(TaMza — NaFig) nV 00) Tuan — Brod), VE) = Vered Tuas = Mol’) —ieLCC\ TM — Brod), where I(Po) and V(r) are the corresponding mode current and voltage at any other point re an } (88) Jon dain, Iu = Taird, ee EI] ype [As for the case of the uniform line the voltage-current relations may be schematically represented by a transmission-line diagram similar to that of Fig. 1-2 whieh indicates the choice of positive directions of V and I Gf2is replaced by r). The relations given in Eqs, (98) between the mode yoltage and current at two points on a spherical E-line may be rephrased jn impedance terms. On introduction of the relative impedance at any point r, 20 = Te) ‘which is counted postive in the direction of increasing r, and division of Eqs. (08), one has it Zire) cto) (99) key) + 52") ” 20 Sec, 18] NONUNIFORM SPHERICAL WAVEGUIDES 53 where x = kr, y = kro and FW)NMe) — No)I@) _ 1 one) FLOR.) = Roig) ~ LED NAG) — MAT) Outed = HRW = Botan ~ TEA (09a) cay) = DBD) = RTD) = gy Hew) = Fan) — Medi) — eta(zy) foley) Ctaly.n), ‘The functions ct, and Ct, are termed the small and large spherical cotangents, respectively; their inverses, tn, and Th,, are correspondingly called small and large spherical tangents. As in the case of the radial functions of Sec. 1-7 this nomenclature is based on the asymptotic identity of the spherical and trigonometric functions at large kr. Plots of the spherical functions are not as yet available.t Equations (99) apply to any E-mode, For conical regions of the type indicated in Fig. 1-166 the lowest: mode is transverse electromagnetic and n= 0. For this mode ct.(zy) = Ctley) = cot(y ~ 2), zy) = and therefore the transmission-line description reduces to that of @ uniform line with a propagation wave number # and characteristic impedance f. From Eqs. (94) and (95) it is evident that the transmission-line description of an H-mode follows from that: of an B-mode on the duality replacements of I, V, ¢ of the latter by V, I, 1 of the former. Con- sequently the relation between the relative admittances at two points on an H-mode spherical line is given by Eq. (99) with 2’(r) replaced by Y'G). Tn the case of H-modes n is always unequal to zero. “The scattering description of the nth mode on a pherealtranamiasion Vine is based on the spherical Hankel function solutions AY (er) = Fa(hr) + GR aC), AP (br) = Jallr) — 38 kr) (0) of the spherical wave equations (96). ‘The former solution represents fan ingoing and the latter an outgoing traveling wave. For the caso of an E-mode the solution of Eq. (96a) for the mode current may be written, T(0) = TaeBt2 (br) + Tne P(E), (1012) 16f. P. R. Desikachar, “Impedance Relations in Spherical Transmission Lines," Master's thesis, Polytechnic Institute of Brooklyn, (1948).5A TRANSMISSION LINBS [Seo 18 ‘and hence <5eV(7) = Lol (Er) + Leal” (hr), (01) where the prime denotes the derivative with respect to the argument. ‘The generator, or exciting source, is assumed to be at small x. To ‘emphasize the analogy with exponential functions the amplitude hy and phase ja of the spherical Hankel functions are defined by Ba) = (—ar hae, AD) = (43th, } (a2) ‘where ° ROEM), f= (n+ 5 + ton FA) VI@+M@, i= (+05 +t KO A current reflection coefficient 140) = FEE (nye Fatenen 08) can then be employed to characterize the field conditions at the point r. ‘The relation between the current reflection coefficients at the point r and re follows as Tir) = Tulraderttetebt (104) ‘Hence, with the knowledge of tables’ of the spherical amplitude and phase functions h, and 4, a scattering description can be developed in exact analogy with that of the unifprm and radial lines (cf. Sees. 1-4 and 1-7). In fact for an n = 0.Z-modo in the conical guide of Fig. 1-155 ho= 1 and fio = hr, and therefore the to seattering descriptions are identical. A similar scattering description can, of eourse, be developed for the H-modes on a spherical transmission Tine, + ables of the amplitude 1nd phase functions ean be found in did, Aleo of Morse, Lawan, Feshbach, and Lex, "“Seattering. and Radiation from Spheres,” O8RD report reprinted by US, Navy Dept., Office of Research and Tnventions, Washington, D.C, 194¢, The amplitude functions in this report are defined somo- ‘what differonty fYom thoso in Eqs. (102). CHAPTER 2 TRANSMISSION-LINE MODES 24, Mode Characteristics —As outlined in the preceding. chapter the description of the electromagnetic fields within a waveguide ean be reformulated in terms of the voltage and current amplitudes of a cet of mode functions e; indicative of the possible transverse field distributions in the waveguide. ‘The resulting transmission-line description, though formally independent of the form of the mode functions, depends quanti- tatively on the characteristic impedance and propagation constant of the individual modes. In many cases these two fundamental mode charac- teristics are simply interrelated so that a knowledge of only the mode propagation constant is necessary. For the case of waveguides with ‘walls of finite conductivity the mode propagation constant = a ++ jf is complex (¢f. Seo. 1-6). The attenuation constant « and the wave- number 8 depend upon the cross-sectional dimensions, the conductivity, and the excitation wavelength 2 of the given waveguide, ‘These mode characteristics must be known explicitly for quantitative transmission- line considerations. Their computation requires a knowledge of the mode field distribution or, equivalently, of the mode function. In this chapter the explicit form of tho electric and magnetic ficld distribution in the various modes of, equivalently, of the vector mode funetions e; will be presented for several uniform and nonuniform wave- guides, ‘The customary engineering assumption of exp (jet) forthe time dependence of the fields, with suppression of the time factor exp (fat) adhered to. Electric and magnetic field intensities are expressed a3 rms quantities. Concomitantly Vand Js, the voltage and current amplitudes of the normalized mode functions ey are rms quantities. As a conse- quence the total mode power flow in the transmission direction is given by Re (Vif). First-order values of the attenuation constant a and the cutoff wavelength Xe will be stated for modes in several types of wave- guide. In addition maximum electric-feld intensities, power expressions, te,, will be indicated in several eases. Although the principal concern of this chapter is to provide quantitative data for transmission-line computations in waveguides of different cross sections, the presentation includes the requisite mode information for the theoretical computation of the equivalent circuit parameters of waveguide discontinuities (cf. See. 35). 5586 ‘RANSMISSION-LINE MODBS [Seo 22 Plots of the clecirie- and magnetic-feld distributions are desirable ‘as an aid to the visualization of the field distributions in the various modes, Mode pattems of two types are useful in this connection. ‘The ‘one type indicates the electric- and magnetic-field strengths on transverse and longitudinal plenes within a waveguide; the other shows the magnetic-field intensity or, equivalently, the current density on the inner surface of the guide. ‘The former of these patterns readily furnishes qualitative information as to the location of points of maximum field strength, power flow, ete, The latter yields information on current flow and is desirable in connection with questions of dissipation and of coupling by apertures in the guide walls. Since a number of mode patterns will be presented in this chapter, it is desirable to say 3 few words as to the construction of the patterns. Each pattern depicts the instantaneous field distribution in a traveling ‘waveguide mode. For a given mode it is desirable to indicate quantita- tively the intensity of the electric- and magnetic-field distribution on a specified plane. Such information can be portrayed in the usual type of flux plots only if the Jeld lines are divergenceless on the given plane, the intensity of the field being then indicated by the density of the field lines. ‘Mode fields gencrally are not divergenceless in a given viewing plane since field lines generally leave and enter the plane. As a result many of the mode patterns are not true flux plots and hence do not indicate the field intensity everywhere. Nevertheless, wherever possible the density of the field lines has been drawn so as to represent. the field intensity. For example, this coavention has been adhered to in regions where no lines enter or leave the viewing plane. ‘These regions are generally apparent from a comparison of the field distribution in the various sectional views. ‘The mode patteris are drawn so that the relative scale of different ‘views is correct as is algo the direction of the field lines. The following conventions have been adhered to: 1, Bleotrie field lines are solid 2. Magnetic field lines are short dashes. 3, Lines of electri current flow are long dashes. Lines of zero intensity have generally been omitted from the mode patterns for the sake of clarity. ‘The location of these omissions should be apparent and taken into account, to preserve the flux plot. 22, Rectangular Waveguides. a. E-modes—A uniform waveguide of rectangular cross section is deseribed by the cartesian coordinate aystem xye shown it Fig. 21. ‘Transmission is along the = direction, Tn a rectangular waveguide of inner dimensions a and b, the B-mode Bro, 22] RECTANGULAR WAVEGUIDES 87 functions e{ normalized over the cross section in the sense of Eqs. (1-5) ‘re derivable from the scalar functions a = 2 sin Me sin Fy, a) © LO ute fms > + nt By Eqs. (1°3), (1-10), and (1) the field ode is given as where m,n = 1, 2 3, tribution of the (20) (2) where the field variation along the z direction is determined by the ‘transmission-line behavior of the mode voltage Vi(2) and current iz) [of. Bq. (1-11)]. The components of the ‘orthonormal vector e{ are given directly by B. and B, on omission of | the amplitude factor Vi. ‘The culo wavelength for the 1 Bnw-mode is y | a ot ,- Aye.) Vebewd obit ser From Eq, (1°50) the Hqa-mode attenuation constant due to dissipation in the guide walls is mt + nt 1 @58 TRANSMISSION-LINE MODES [820.22 where the characteristic resistance 0t is a measure of the conductivity properties of the metal walls (of. ‘Table 1-2). ‘The attenuation constant, is a minimum at the wavelength ) = 0.577X,. ‘The maximum value of the transverse ris electrie-feld intensity of the Hag-mode is (5a) (5b) Since an E-mode possesses a longitudinal component of electric field, the ‘maximum of the total electric field is dependent on the impedance condi- tions in the guide For a matched nondissipative guide the total average power flow along the positive 2 direction is given by Pi = Re (Vii) © In Fig. 22 are pertrayed the field distributions of the Bu, Hz, and Ezemodes in a rectangular guide of dimensions a/b = 2.25 and excitation such that 4,/a = 1.4. The mode patterns on the left-hand side of the figure depict the clectrie and magnetic lines within transverse and longitudinal sections of the guide. ‘The right-hand pattems show the ‘magnetic field and current lines on the inner surfaces at the top and side of the guide , H-modes—The H-mode functions e/, normalized over the cross- section, are derivable from scalar fun @ where 1 Pas VW net t Hi ti ‘Fio, 22—Field distribution for Esmades in ostanguler guides 1, Cross-sectional view 2 Longidinal vow 3. Sunface view60 TRANSMISSION-LINB MODES (Sne. 22 ‘The field components of the Haq-mode follow from Egg. (1-4), (1-10), and (7) as (8a) H, = 1 (88) The z dependence of the field components is determined by the trans- mission-line behavier of V/(@) and 1/() (ef. Eq. (1'11)}. On omission of the amplitude faxtor V{ the components of the orthonormal vector function ef are given directly by E, and 2. ‘The cutoff wavelength of the Hn,-mode is we ova . ey ae ae bens and is exactly the same as for the Emy-mode. The attenuation constant for a propagating Hmwmode due to dissi- pation in the guide walls is where the character’stie resistance & (¢f. Table 1-2) is a measure of the conductivity of the metal walls. ‘The maximum electrie-field intensity in a Hagmode is for m x 0, n#0 ®) 2h) ab mW na" mb fom (11a) Sve. 22] RECTANGULAR WAVEGUIDES 6 and occurs at values of 2 and y for which tan Me = £2 tan My For m x0, n ~ 0, the maximum electro fold i (18) and occurs at integral multiples of x = a/2m. For m = 0, n x0 the maximum field has the same magnitude es in Eq. (118) but occurs at integral multiples of y = b/2n. Tn terms of the rms mode voltage ¥/' the total power carried by a ‘traveling Hns-Wwave in a nondissipative guide is given by Penna (3) er. 2) ‘The Hirmode is the dominant mode in rectangular guide and hence will be considered in some detail. Instead of the voltage V\, and current Ti employed in Eqs. (8) voltage V and current J more closely related to low-frequency definitions ean be introduced by the transformations vow fey t= In terms of V and J the nonvanishing field components of the Hirmode V ent, B, =~ FsinZr, as) For guide walls of finite conductivity the attenuation constant of the meas a2) oe ;2 TRANSMISSION-LINE. MODES (Se0.22 ( to a first order. ‘The maximum eleciric-field intensity occurs at. the center of the guide and has the magnitude |V|/5. ‘The rms voltage V in ‘a traveling wave is related to the total power flow by Ive and the guide wavelength is » (ase) or a nondissipative guide ‘Mode patterns for traveling Hy, Hs, H=.-modesin rectangular guides are displayed in Fig 23. As in the case of Fig. 22, a/b = 2.25 and dy/a = 14. ‘The mate patterns on the left indicate the field distribution ‘on cross-sectional and longitudinal planes; those on the right depict the lectric-current distribution on the inner surface of the guide. ¢. Modes in a Parallel Plate Guide. ‘The modes in a parallel plate guide of height & may be regarded as appropriate limiting forms of modes in a rectangular guide of height } as the width a of the Iatter becomes infinite, As noted in Secs. 22a and b, the modes in a rectangular guide of height 6 and width aform a discrete set. However, as the width of the rectangular guide becomes infinite the corresponding set of modes assumes both a discrete and continuous character; the mode index m characteristic of the mode variation along the finite y dimension is dis- crete, whereas the index m characteristic of the variation along the infinite x dimension becomes continuous. ‘The complete representation of a general field in 9 paraliel plate guide requires both the discrete and continuous modes. Tor simplicity, we shall consider only those discrete modes required for the representation of fields having no variation in the x direction. For the representation of more general fields, reference should be made to See. 26 wherein a typical representation in terms of continuous modes is rresented. ‘The discrete F-modes in a parallel plate guide of hei from the scalar functions ht b are derivable sin SY ange oxy Bo, (164) E.=0, He = 0, =P Hak (160) where as above the 2 dependence of the mode fields is determined from the transmission-line behavior of the mode amplitudes V/’ and I. The components of the orthonormal vector ef’ are obtained from Z, and Ky ‘on omission of the mode amplitude VJ. The cutoff wavelength Ni of the Hoy-mode is 2b/n, the same as for the Esymode. ‘The attenuation constant of the Hex-mode due to finite conductivity of the guide plates is aa_(3) Ni (166) ic resistance @ of the metallic plates may66 ‘TRANSMISSION-LINE MODES (Seo, 28 be obtained from ‘Table 1-2. ‘The decrease of the attenuation constant ‘vith increasing frequency is to be noted. 2.3, Circular Waveguides. a, B-mades—A uniform waveguide of circular cross section is most conveniently described by a polar coordinate system re as shown in Tig, 24 For a cross soction of radius @ the E-mode vector funetions e{ normalized over the cross see- tion in accordance with Eqs. (1-5) are derivable from the scalar functions Inf - 28 * ¥ xTmintas) sin ™% “on where m= 0,1, 23, , . ia= 1) iim =0, Pail em en a2 mao, and x: = Xan, the nth nonvanishing root of the mth-order Bessel function J4(x), is tabulated in Table 211 for several values of m and ‘Tamim 24.—Roors oF Jala) = 0 tyre (ind — 1)@nt ~ 31) an (420-2) 5m ee DP Nie Lt 12 a fs fo] az 1 | 2.405] 3.si2| 5.130 7.583 | 8.771 | 9,936 | 11.086, 2 | 5.320] 7.016} 8.417 11.055 | 12.380 | 18.580 | 14.891 3 | s.6s4| 10.173 | 11-020 14.872 4 | an-702 | 13.323 | 14.795 On use of Kgs. (1-8), (1-10), and (17) the field components of the Bayemode become ir polar coordinates a il ea f me, fat See. 23) CIRCULAR WAVEGUIDES er H,= Ft sty 5) --1 cos Me opie sin H, = 0. ‘The ¢ variation of the mode fields is determined by the transmission-line behavior of V/and Ij given in Eqs. (1-11). As is evident from Eqs. (18) the Znwmode (m + 0) is degenerate and consists of two modes with even or odd angular dependence. ‘Though not explicitly shown each of these modes is, of course, characterized by a different mode voltage and current. ‘The polar components of the vector e/ are obtained from E, and Ey on comission of the amplitude Vj, ‘The cutoff wavelength of the degenerate By-mode is, 2n Me = x (19) where the roots xnx are given in ‘Table 2-1. ‘The B-mode attenuation constant due to Snite conductivity of the guide walls i (20) where the frequency-dependent characteristic resistance & of the metal walls may be obtained from Table 1-2. ‘In terms of the rms voltage V; the total power carried by an Emwemode ey Mode patterns of the instantaneous field distribution in_ traveling waves of the Fo, Bir, and Barmodes are shown in Fig. 25. The excitation frequency is such that 4,/a = 4.2. The left-hand views of Fig. 2 depict the electrio- and magnetic-field intensities on transverse fand longitudinal planes in which the radial electric fied is maximum, The right-hand view shows a development of the magnetic-ield and cleciric-current distribution on half the guide cireumference,68 TRANSMISSION-LINE MODES (Seo. 28 Bec, 239] CIRCULAR WAVEGUIDES 69 b. H-modes—The e{' vector mode functions normalized over the circular eross section are derivable from sealar functions no Egg le) m= 0,1, 2,3, mb (2) where and x{ = xt is tho nth nonvanishing root of the derivative of the mth- order Bessel function. Several of the lower-order roots are given in Table 22. Tanue 22:—Roors or Ji(x’) = 0 : 2 dnt 3 Tmt + 325m" —9 xa = (nt 28- iim TaD 7 y om>o 3.054 | 4.201 | 5.317 6.706 | 8.015 | 9.282 9069 | 11346 | 12 682 13.170 6.416 | 7.501 | 8.578 10.520 | 11.735 | 121932 13.987 From Eqs. (1-4), (1-10), and (22) the field components of an Hme-mode are found to be t ) ag (23a) B= 0, p (xin te Bh ot 2) “a(x) sin (230) Tio, 25—Feld distribution for B-modes in cbculae waveguide 1 Crosson iy 2, Sunfece view fom mg.70 TRANSMISSION-LINE MODES (sro. 28 As evident from Eqs. (23) the Hax-mode (m # 0) is degenerate; ie., there ‘exist two Hwn-modes, one of odd and one of even angular dependence. ‘The polar components of the normalized mode vectors e'' are obtained from B and Hon omission of the 2 dependent voltage amplitudes Vf ‘The cutoff wavelength of the degenerate Hma-mode is = 2 Me a” (24) ‘the roots xi being given in Table 2-2. Except for the degenerate case m = 0 the cutoff wavelengths differ from those of the Zarmodes. ‘The Hns attenustion constant due to dissipation in the guide walls is wen co) resistance @ of the metal walls ‘The frequency-dependent. charucter may be obtained from Table 1: “The total average power carsied by a traveling Hwmode in a matched nondissipative guide is expressed in terms of the mode voltage V/ as ai) er 5) ‘The dominant mode in circular guide is the Hy. In a nondissipative guide the wavelength of propagation of the Hy-mode is, (2) Sala, ‘The maximum rms slectrio-field intensity of the Hurmode oocurs at the axis of the guide anu! has a magaitude wt 130 ey Frau (28) In terms of Hass the maximum average power carried by the H-mode is Pou 87 x10 t= (5h) where all units are MIKS. Mode patterns of the instantaneous field distribution in traveling Hu, Hyy, and Ha: -modes are shown in Fig. 20 for 2j/a = 4.2. The transverse and longitudinal views are in the plane of the maximum OB, (29) See. 23] CIRCULAR WAVEGUIDES a Fro, 20-—Fald distribution for H-modes in circular waveguide. 1. Cromseetional view 2, Longitadinal view through plane i 13, Suriaea view from2 ‘TRANSMISSION-LINE. MODFS (Seo. 24 radial electric field, The views on the right-hand side of Fig. 20 depict the magnetic-ield and electrie-current distribution on half the guide circuraference. 24. Comxial Waveguides. 0. E-modes—The deseription of the uniform coaxial guide depicted in Fig. 2°7 is closely related to that of the circular suide considered in the previous section. ‘The coaxial guide is described by a polar coordinate réz system in which the outer and inner conductors are at radii @ and b, respectively, and the transmission direc- tion is along the z-axis. ‘The vector functions ef normalized in accordance with Eqs. (1-5) and character- istic of the H-mode fields are derivable [e. Eqs. (18)] a8 gradients of scalar functions. ‘The scalar function appropriate to the lowest E-mode in a coaxial guide is 9 = BE 0) Fro, 27.—Coanial waveguide cross Arr ing Henee by Eqs. (1-10), the field components of this mode are By = Vio h, spring , 1 , Hyatt k en print Epa Be= HH, This transverse elestromagnetic or TEM-mode is the dominant, or principal, mode in coaxial guide. Its cutoff wavelength is infinite, and hence \, = 2 for the dominant mode. ‘A. more customery definition for the voltage and current of the principal coaxial mode is obtained by multiplication and division, respectively, of the normalized voltage and current in Eq. (31) by ling oN a See, 24] COAXIAL WAVEGUIDES 3 Tn Lerms of the new voltage V and current I the nonvanishing components of the dominant mode become 2) ae ‘The maximum eletre-feld intensity in the principal mode occurs at the surface of the inner conductor and is of magnitude Bow = HL. 3) bind For constant outer radius and voltage the maximum electrie-ield intensity is a minimum when a/> = 2.72. Ona matched coaxial line the rs voltage V is related to the total average power flow P by pay ive 4) nf For constant outer radius and |aa|, the power flow is a maximum when a/b = 1.85. The attenuation constant of the dominant mode due to dissipation in the inner and outer conductors i @ +e), G5) . ane where Gi, and & (ef, Table 1-2) are the characteristic resistances of the metals of which the outer and inner conductors are constituted. For fixed outer radius and wavelength the attenuation constant is a minimum. when a/b = 3.6 provided Gis = Gt. ‘The normalized ef vector functions characteristic of the higher B-modes are derivable from the scalar functions Ba tal x ) nm (36) where Vee Ilx ) Nala) = re ve Ne Fax * (63 -] 01,23," 1 itm =o, 2 ifm.m TRANSMISSION-LINE MODBS (Se. 24 ‘Phe quantity xi = xmu is the nth nonvanishing root of the Bessel- ‘Neumann combination Z(cx;), whereé'c = a/b. |The quantities (¢ — 1)xx are tabulated in Table 23 as « function of the ratio efor several values of mand m ‘ass 234—Roore oF Jn(OXNn(a) ~ Nafex)a(a) = 0 ‘Tabilatad inthe form fe — xan» > Sy nu fom fos fo | wm | om | oe 10 [su2| 3.2] 3.142] 3.142) 6.250 | 6.28) 0.280 | 0.280 ia faaa| ana | alis7| sitet | oss | olasi| e280 | 6:20 12 ago] aiae| aio: | ai1s7| 6.982 | o.2s5| oan | 6.200 1a faciay| ats | ats | 3:20] sas | eias7| ean | e.aai 1 [auar| ass | 2.208 | 3'a0| sast | 6:20] oar | 6.382 us | .s5| 3.101 | 3.257] 3:96 | 6.250] 6.203 6.ax2 | 8.307 us [ass | 316s | 3.27 | asa | s.2| 6.00] oso | Sear v's [auss|ais| sa | sie | save| sao] easr| sees 20 [aiss|aiv|aa | 37 | som| saz) 6a | oer 2's | sino asa 6.200 | 6.95 6 a0 | a.oor | 3am 6.358 | 6.357 a's | a.oes | 3.200 8.230 | 0.381 0 | str | ase f2i9 | ana Seyejefef eps) a | oo 10 | 9.495 | 9.495 | 9.425 | 9.425 | 12.500 | 12.008 | 32.500 | 12.500 11 | 91498 | 0.45 | 9-497 | o°4a0 | 19060 | 12.507 | 19.508 | 12-500 12 | 9's | 9°46 | 9491 | ocaso | 12-060 | 12°s07 | 12.3 | a2 878 13 | ocdai | o.ao7 | 949s | ovasr | 12-300 | 12-508 | 32.77 | 12-000 rat | 9.29 | 9:0 | o-4er | ocens | 12.865 | 19.570 | 32.589 | 32°008, us | 9.2 | acs | 9.496 | 91502 | 12365 | 12.571 | 32.901 | 12-024 16 | ou4a2 | 9.404 | 9.460 | 9°50 | 12-004 | a2.sra | 12.000 | x2-c48 1's | 9°4an | as | 9°40s | 9'5er | 12-560 | 12-577 | 19-010 | 1080 2.0 | o.ais| acs | 91528 | 0:62 | 12.961 | 12 581 | 12.000 | 12.738, 2's oa | ogo]... | 9-89 | 2.958 | a2 508]... | 12.874 3.0 | 94s | oa 10.0" | 12.583 | 12.008 i302 3's | 9.402 | alana 10.2 | 12.000 | 2-819 12 4.0 | 9am | aan 12-645 | 12.681 13 GF HB Duis, “Vata a Ro for RatatPraqwed io Con Gavin Jou, Ba rye FY.) 4-09 04, ‘The cutof wavelength of the Bu, of the tabulated values (¢ — I)xnn a8 -mode may be expressed in terms 2a) 2 7 M= Glia 7 ne4j2305+. (87) See. 24) coaxraL WAVEGUIDES 5 ‘The field components of the Bqy-mode follow from Eqs. (1-8), (1-10), and (36) as B= vi tax A) nm By= £V Em (x ) a ms, (38a) De Xe Bo ir 2, Hy = #1™ Za(xcf) 5 me, as 7p (2) 008 aso) y= 18 22.(xa) 20 mo, H=0 From the form of Eqs. (38) it is apparent that the P..-mode (m > 0) is degenerate and may have either of two possible polarizations, each distinguished by a different voltage and current amplitude. The polar ‘components of the ef vector are obtained from Hq. (88a) on omission of the z-dependent voltage amplitude Vj. ‘The attenuation constant due to finite conductivity of the inner and outer conductor is a Jala 4 a-|4 Falexs) 1 @0) Frlexd ~* where @. and 64 (of. Table 1-2) are the characteristic resistances of the inner and outer conductors, respectively. Tn terms of the rms voltage V the total power transported by a traveling Bqy-mode in s matched nondissipative g (40) ‘The instantaneous field distribution in traveling waves of the Hey, Ey, Bu type are shown in Fig. 28, ‘The mode patterns are all drawn for the case a/b = 3 and d,/a = 4.24. The left-hand views portray the electric and magnetic field distributions in the transverse and longi- tudinal planes on which the radial electric field is a maximum. ‘The right-hand patterns show developed views of the magnetic-field and clectric-current distribution on half the circumference of the outer conductor.76 ‘TRANSMISSION-LINE MODES [Be0, 24 my I | apie rt sre.24) | COAXIAL WAVEGUIDES 7 b, H-modes—The vector functions e’ normalized in accordance with ‘gs. (1'5) are derivable from scalar functions we a(x De om a where (ut) Jaxif) Nata = "a i) sa000 2 (Tee, (=) eel (1 —(25)'| [1 -(% (eel b- & & m=0,1,2,3,-°°, and xi = xbq is the nth root of the derivative of the Bessel-Neumann combination Zm(cx!) with ¢ = a/b. For n = 1 the quantities (¢ ++ I)xsa are tabulated as a function of the ratio ¢ in Table 2:4; for n> 1 the quantities (c — 1)xme are tabulated in Table 2:5, Tan 24.—Finsr Roor oF J(ex)NWG) — Wslex"Wa0e) = 0 ‘Tabulated in the form (e+ Tubs, > 0) T Se pomp a 10 2.000 4.000 0.000 ui 2.001 ‘on 002 12 2.002 £000 0.008 13 2-000 ou so ia 2-009 015 solr Us 2013 £00 6.018 16 2.018 4025 oon 18 2.0m 4000 5.086 20 2081 so 5.07 28 2048 3.080 5.781 x0 2.056 3.008 5.082 35 2087 ae Sas 43 2.05 | 8.700 5.240 Tho, 28—Fidd datibution for E-modes in consal waved 1, Crosmeestional view 2 Longitudinal iow Yieoush plane Uf 4 Surface view from er ‘In terms of the tabulated values, the cutoff wavelength of an Hmi- modi can be expressed as 2 Tp, th Fe (420) H+) form =1,2,3, Nee and for an Hay-mode as (420)78 ‘RANSMISSION-LINE MODES (Seo. 24 It is evident from these equations that the Hi-mode is the dominant H-mode in coaxial guide, The eutofT wavelength of the Hs-mode is identical with that of the Hy;-mode, ie, xh, = xn and ean be obtained from Table 2°. ‘Tene 25.—thoune Roors oF Info Nala’) — Nelo tx! Tabulated in the form fe = Vem (n> 1) le ]ol = : Seo. 24] COAXIAL WAVEGUIDES 79 Ry eH L873 F(a} 223 10 [a.u2{s.ua| a.12| 3.02] 0.250] 6.280 ri [sius} sma] 3.148 | 3356] 62st] 6:2se 2 va [3.45] ax | ater] 3103 | 6/285 | 6.288 rs [5.0] ao | s.1s| 3:20 | 0.287} 6.203) ! ra [3.185] sar] 3.200] 3'a10| 0.200] 6.20 1s | aust] sass] ar | 30 | oma] 6.305 re [ser] 35] 3's | 39 | 6.206] 6.35 18 | s.1s2 smi} 34 | a7 | 6.333 20 | aue7| sa | 3.5 0.353 2s | 3.205| 390] sao | My a0 [aan | sae ara | a a5 | 3/305 | 3.00 oes #0 | 9.396 | 3.758 eee a “| uf ox | x w | 2 | a 11 [91425 | 9-126] 9-427 | o°an0 | 12:67 | 12\567 | 121508 | 19 570 12 [91426 | 9.428 | 9-453 | o-4se | 19.567 | 12-800 | 32°s0o | 12.870 | 13 [o.aa7| oat] 9.442 | o-so1 | 12.566 | 12.571 | 121570 | 12 500 La | 94a faces | olase | olase | ta'sro | 12.374 | 12°50 | 12-611 ; 1.6 | 94a. | 9.40 | 9.407 | 9.511 | 12\571 | 12.578 | 12.598 | 12.681 | Uo | ovis | ova | ovssz | o\sit | 49.070 | 12.960 | 19.000 | 1a.oat | 1s | 91430 | 0.458] 9-515 | 9°600 | 12.577 | 12.501 | 32%6a4 | 12-704 zo |o.s| ova | 9.852) 9:58 | 12\561 | 12\001 | 19.001 | 12-701, Lo * 25 | 0.460] 9:5 | 9.665 | 9.000 12.008 | 12.000 | 12.700 | 12.02 in TAF sa |o.rs| 0.80] 0.77 | 10.1 | 12-605] 12.080 | 12.82 | 13.00 a5 joss | 9:50] 9-89 12619 | 12.002 | 12.01 | 13.2 40 | 9:900 | 9.655 | 10:0 w2iea1 | 12.725 [19.0 | 13.8 ‘oa Fock doit ova Der ana =, Fie era 7 mn 90) ae anh Ti Use ant api mponte oo Hosmote aa given by Eqs. (1-4), (1°10), and (41) as Bon a7! 24 (sf) me . Xb gt (4 7) 008 (43a) 5-8 (9) SS ma, E.=0, ' : . HCH 4 Hey ct ea EE CAHIR) Cosh 1 comet ae80 TRANSMISSION-LINE MODES (Bre. 25 Xan (yp) 008 Fax ) sin ™4 He= 217 2 (x2 com (432) dyn xt 7) cos H, As in the case of the B-modes, the Haq-mode (m > 0) may possess either of two polarizations, each characterized by a different, vollage and current. ‘The polar components of the normalized ef vectors are obtained from E, and 2, of Eqs. (43a) on omission of the z-depondent amplitude V/. For a traveling Hymode the attenuation constant due to finite conductivity of the guide walls is given by (Se5@) +P] Ge) +[% Tad 4 os] QUN? a Tulexd FVi= Ont Tr mab -@)I-b-@)] where Gand @, are the charactoristic resistances of the metals of which the outer and inner sonduetors are composed (ef. Table 1 ‘The total average power carried by a traveling Hm-mode in a matched nondissipative coaxial guide is expressed in terms of the rms voltage Vas, > (44) Pes me) Ive 4s) Mode patterns of the instantaneous field distribution in traveling waves of the His, Ha, and Hs, coaxial modes are shown in Fig. 29. The patterns portain to coaxial guides with a/b = 3 and 2,/a = 4.24. The patterns on the left depict the electric and magnetic field distribution in transverse and longitudinal planes on which the transverse electric field is a maximum; those on the right portray the magnetic-feld and electrie- current distribution on half the circumference of the outer conductor. 246. Elliptical Waveguides.'—An elliptical waveguide is » uniform region in which the transverse cross section is of elliptical form. As Of.L. J. Chu, “Bleetromagnetic Waves in EUiptio Hollow Pipes,” Jour, Applied Phys, 8 September, 1758, Siatton, Morse, Chu, Hutner, Bliptic Cylinder and ‘Spheroidal Wevefunctions, Wiley, 194, S00. 25) ELLIPTICAL WAVEGUIDES at illustrated in Fig. 2-10 elliptic coordinates fy (the coordinate 7 is not to ‘be confused with the free-space admittance » employed elsewhere in this, volume) deseribe the eross section and the coordinate z the transmission direction. ‘The rectangular coordinates zy of the cross section are related to the coordinates of the confocal ellipse § and confoval hyperbola n by x = q cosht cen | (6s) y= qsinbé siny, where 2q is the focal distance. ‘The boundary ellipse is defined by the coordinate § = a with major axis, minor axis, and eccentricity given by 2 cosh a, 2y sinh a, and e = Leosh a, respectively. ' The ease y of a circular boundary is described by e = 9 = 0 with g/e finite ‘The mode functions e! charac- teristic of the E-modes are deri able from sealar functions of the form she = Reale) Sealer), (Aa) = RonlExdSOm(Hn0), (70) where m= 0, 1, 2, 3, ++ + and 2X = ton is the nth nonvanishing. oot of the even mth-order radial Mathieu function Req(a,x), whereas ox; ™ eqn 38 the nth nonvanishing root of the odd mth-order redial Mathiew function Roa(a,x). ‘The functions Sen(n,x) and Som(n,x) are even and odd angular Mathieu functions. In the limit of small x = kag ‘the Mathieu functions degenerate into circular functions as follows: lim Sen(n,x) = eos m¢, lim Soa(nx) = sin mg, (as) Tim Rea(ex) = lim Roa(tx) = aff Jako), and, correspondingly, the confoeal coordinates f and y become the polar coordinates r and ¢. The field components of the even ,Emq-tnode follow from Eqs. (1-3), (1-10), and (47a) as 7 Rela Grd Sem.) Be . q-Veosh’ 7 By = — Vj Reale, (49) “a Veosh* & — cosy Bem -B RM MRealbaxd Seale,82 (See. 25 He (498) "The prime denotes -he derivative with respect to €or n, and Ji, = 2x/Ne denotes the eutof wavenumber of the even mn mode. The # dependence of tho fiolds s determined by the transmission-line behavior of the mode amplitudes Vj and 1}, The odd Bwrmode is represented in the same manner as in gs (49) save for the replacement ofthe even Mathieu fune- tions by the odd. The components of the mode functions e/ are obtained from B and B, of Eqs. (49a) on omission of the amplitude Vand insertion of a normalization factor. ‘The e/-mode functions characteristic of the H-modes in eliptical sides are derivable from sealar functions of the form Ms = Rent xi) Sea4eX9, (602) 0 = Ronl Ex) S0m( 0x0 os) where mm = 0, 1, 2,3; = = and xf = athy and ox nonvanishing roots of the derivatives ofthe radial Mathiet functions as defined by Ree (ex) Roz (ayox,) ‘The field components of the even Haq-mode are given by Eqs. (1-4), (1-10), and (50a) a8 (51) 1 Rend fax) Se opr Regltva (62) E,=0, 1 Rea buat) SemC rex) Hy= =I . g Veosh* Hy = m1 Beakbu d. (53) de pn ‘ ) Ha = i Hap een) Smo. ‘The icld components ofthe odd oH/mw-modeare obtained from Eas. (52)and (68) on replacement of the even Mathieu functions by the odd. ‘The com- ponents of the of mode funetions are obtained from Hy and J, on omis- sion of the mode anplitude ¥/' and eddition of « normalization factor. Src. 25] BLLIPTIGAL WAVEGUIDES 83 ‘The cutoff wavelengths of the Z- and H-modes may be expressed in terms of the roots x: and the semifocal distance g as 2rq, da = 28, xi oe oon omission of the various mode designations. An alternative expression in terms of the eccentricity ¢ of the boundary ellipse is obtained by use of the elliptic integral formula for the eixcumference 1 [vis eearia = YH (5) of the boundary ellipse. ‘The ratio 2./s is plotted vs. ¢ in Fig. 2-11 for several of the even and odd E-and H-modes of largest cutoff wavelengths. sews vn wet wos eos Fro. 211.—-Cutof wavelength of ellipteal wavegulde, It is evident from the figure that the .Ay-mode is the dominant mode in an elliptical waveguide. The splitting of the degenerate modes (m > 0, (0) of a circular guide into even and odd modes is also evident from this figure. Computation of power flow and attenuation in elliptical guides involves numerical integration of the Mathieu functions over the guide cross section. The reader is referred to Chu’s paper quoted above for quantitative information, ‘Mode patterns of the transverse clectrie and magnetic field distribu- tion of several of the lower modes in an elliptical guide of eccentricity = 0.75 are shown in Fig. 212, The patterns are for the Hor, Hu, Hy, Boy, Bus, and oEux -modes.Pio, 212—Field distribution of modes in llipteal waveguide. Cross-sectional view: 2.6, Space as a Uniform Waveguide. a. Fields in Free Space—Free space may be regarded as a uniform waveguide having infinite cross- sectional dimensions, A transmission-line description of the fields within free space can be developed in a manner similar to that of the preceding sections. The cros-seetional directions will be described by the zy coordinates and the transmission direetion by the ¢ coordinate of a rec- tangular coordinate system as shown in Fig. 213. In this coordinate system the general clectromagnetic field can be expressed as a superposi~ tion of an infinite set of B- and H-modes closely related to those employed for rectangular guiles (cf. See. 22). In the absence of geometrical structures imposing periodicity requirements on the field, the required modes form a continuous set of plane waves, cach wave being character- See, 26} SPACE AS A UNIFORM WAVEGUIDE 85 ized by a wave number k indicative of the direction of propagation and wavelength. ‘The ef vector functions character ¥, istic of the B-modes are derivable from sealar functions of the form citertted Bah (66) where — 2 k, i.e, when @ is imaginary. ‘The power flow per unit area in the 2 in terms of the rms mode voltage Vi. a8 ireetion can be expressed ravi (61a) oa ivi = shy for the ease of a pronagating B-mode or as Sue, 26) SPACE AS A UNIFORM WAVEQUIDE 87 vet GY tree = acco arte oxo for the case of a propagating Hi-mode. ‘The dominant Z- and H-modes are obtained from Eqs. (67) and (59) by first placing ks = O and then ky = 0, or conversely. ‘These modes are evidently TE] waves with the nonvanishing components Ey = Vie)y Be = Vee), He= -Ii@), Hy = K@) and are seen to be polarized at right angles to one another. ‘The wave- length 2, of propagation in the = direction is equal to the free-space ‘wavelength 2 for these modes. 20 (62) 5 tana nom (ne way) fhe + ‘Woelrath in em Fig. 214.~Atmospherle attenuation of plane waves, (@) Atmosphere composed of 10 tam. of water vapor per cubic meter. (©) Atmosphere composed of 20 per cent oxy at ‘otal peseure of 76 em. He 3s 0 ‘The attenuation constant of plane waves in free space is determined solely by the dielectric losses in the atmosphere of the space. Because of the importance of this type of attenuation at ultrahigh frequencies theoretical curves of the attenuation (8.686), in decibels per kilometer, due to presence of oxygen and water vapor, are shown in Fig. 214 as @88 TRANSMISSION-LINE MODES Bre. 26 b. Field in the Vicinity of Gratings—The analysis of the electro- magnetic field in the vicinity of periodic structures, such os gratings in free space, is in many respects simpler than in the ease where no perio- icity exists, Instead of a continuous infinity of modes, as in See. 26a, a denumerably infinite set of Z- and H-modes is present with only discrete values of k. and fy. When the excitation and geometrical structure is such that the field has a spatial periodicity of period a and 6 in the trans- verse 2 and y directions, respectively, the only permissible values of ky and ky are (63) “tn no A cituation of this sort obtains, for example, when # uniform plane wave falls normally upon x planar grating having a structural periodicity of length a in the direction and bin the y direction. 'The Z,-and H-modes, in such a spave are given by Eqs. (57) and (69) with &, and ky as in Eqg. (G8). The cutoff wavelengths of the By- and Hrmodes are both equal to (64) which is characteristic of the wavelength below which the plane waves are propagating and above which they are damped. On introduetion of the substitution (68) it is appazent that these modes can be interpreted in terms of a disorete set of plane waves (diffracted orders) defined by the angles @ and ¢. In many practical eases the excitation and dimensions fare such that )> a> b, and hence only a single mode, one of the dominant modes shown in Eqs. (62), can be propagated. As a conse- quence the field is almost everywhere described by the voltage and current of this one mode. ‘A similar mode analysis can be applied to describe the fields in the vicinity of a periodie structure when the excitation consists of a plane wave incident at the oblique angles 6’ and ¢’. If the spatial periodicity of the structure is again defined by the periods a and 6 in the z and y directions, the only permissible Z- and H-modes, shown in Bqs. (57) and (69), are those for which ke AM, m= 0, £1, 42,05 ° 5 “ (65) hyo 2B my, nO, £1, 32, See. 27) RADIAL WAVEGUIDES 89 where Kj = k sin cos @! and ki = sin 0’ sin ¢ are the wave numbers of the incident excitation. The cross-sectional wave number of both the Ee and the H-mode equals -2 (@_2ES) G army and as in the previous case these wave numbers characterize a discrete set of plane waves (or diffracted orders) Frequently the excitation and dimensions are such that only the dominant mode (m = n = 0) with kz = 1! and ky = ky is propagating. For example, if ¢’ = 0 and a > 6, this situation obtains, as ean be seen from Eq, (66), when (68) D> a(t + sin 8) = es, i, when the next higher dftraction order does not propagate. Under these conditions the dominant mode voltage and current. describe the field almost everywhere. The components of the e{ vectors are obtained from the E, and Ey components of the mode fields of Eqs. (57a) and (59a) on omission of the tudes, ‘Tho normalization of the e, veetor functions has not tly stated since it depends on whether the mode index ¢ is continuous or diserete. For a continuous index #, e, must be divided by 2r to obtain a vector function normalized over the infite cross setion to a delta function of the form a(k, — k{) (ky — ky). For discrete index i, e, should be divided by +/ab fo obtain, in accordance with Hag. (15), a vector function normalized to unity over a cross section of dimensions a by 2:7, Radial Waveguides. a. Cylindrical Cross Sections—An example of a nonuniform region in which the transverse eross sections are com- plete cylindrical surfaces of height bis provided by the radial wavewuide $ustrated in Fig. 216. In the rée polar coordinate system appropriate of this type, the transverse cross sections are de surfaces and sion is in the direction of the radius r. Radial waveguides are encountered in many of the resonant cavities employed in ultra high- frequency oscillator tubes, filters, et; ree space can also be regarded as aradial waveguide of infinite height. As stated in Sec. 1-7 the transverse tlectromagnetic feld in radial waveguides cannot be represented, in general, as a superposition of transverse vector modes. There exists nly a selar representation thet, for no H, fel s expressible in termas of B-type modes tnd for no B, el, in terms of Hetype modes, "The delta function B(x — 2) in defined hy the conditions that ie integral be “unity if th interval of integration includes the point 2” and be zero otherwise.90 TRANSMISSION-LINE. MODES (Src. 27 General view Side view Tio. 216-—Radial waveguide of eylindreal eros esction. E-type modes,—The field components of an E-type mode in the radial guide of Fig. 2.15 can be represented as By = VE cos ME 2 OF ms, By = EVES EM sin Fe md, (670) By = ~ i Tig Ty MF F gin ™ = 0, Hy = 1; $2 008 2 °° mo, om meek m ome sin, Fe SPV TE ear 8b * 05 where 2 (ty, moe -(F) Gal ina, @=2 ifn so, m=0,123 55, n= 01,23, ‘The # dependence of the E-type modes is determined by the transmission- line behavior of the mode voltage V! and current {. ‘The latter quanti- ties satisfy the transmission-line equations (1-64) with (6s) Bo. 27] RADIAL WAVEGUIDES 91 For either m = 0 or n = 0 the transverse electric and magnetic fields of ‘each mode have only a single component. As is evident from the form of Eqs. (67) there exist two independent Zw.-type modes with different, ¢ polarizations; the amplitudes of these degenerate modes are different from each other although this has not been explicitly indicated. In a radial waveguide the concept of guide wavelength loses its customary significance because of the nonperiodic naturo of the field variation in the transmission direction. Consequently the usual relation between guide wavelength and cutoff wavelength is no longer valid. However, the cutoff wavelength, defined as the wavelength at which X= 0, is still useful as an indication of the “propagating” or *‘non- propagating” character of a mode. For an E-type mode the cutoff VG) +) and its dependence on r indicates that the mode is propagating in those regions for which \ <2, and nonpropagating when \ > Xi,. In terms of the rms mode voltage and current the total outward power flow in an E-type mode is Re (Vi/*). For computations of power on a matched line it should be noted that the input admittance of a matched radial line is not equal to its cbaracteristic admittance (ef. Eq. (1-74)]. ‘The dominant B-type mode in the radial waveguide of Fig. 216 is the m (0 mode and is seen to be a transverse electromagnetic mode. The nonvanishing field components of this THIM-mode follow from Eqs. (67) as (omitting the mode designations) 1 Ns (69) B= ~ 0, 10), (70a) He Fer and the corresponding characteristic impedance and mode constant as and, (700) Frequently the excitation and guide dimensions are such that the domi- nant mode characterizes the field almost everywhere. The total outward power carried by the dominant mode at any point r in a matched non- dissipative radial guide is Der 2 . PE ay VON mm is evident from the power relation and Eqs. (1-74); the function ha(kr)92 TRANSMISSION-LINE MODES (See. 27 is the amplitude of the zeroth-order Hankel function and is defined in Eq, (1-80), H-type moies—The field components of an H-type mode in the radial guide of Fig. 215 are given by B. B (72a) z m6, Hea 11 ain Eo ng, ig te tc Hy 217% cog 2 OH mg, 2 B= “inv $f, cos 2 me, where —— wml (%): ee ee ‘The edependent mode voltage Vand current Zi obey the radial transmission-line equations (1-64) with Daren Kk TE wwe“) -( ‘The existence of two distinct Hratype modes with different ¢ polariza- tions is to be noted. The cutoff wavelength of the Hrxtype mode is identical with that of the Enatype mode and, as in the latter ease, is indicative of regions of propagation and nonpropagation. ‘The total outward power flow in an H-type mode is given in terms of the rms mode voltage and current by Re (V1!) The dominant H-type mode in the radial guide of Fig. 2-16 is the a (73) m=0,n=1 mode, ‘The nonvanishing field components of this mode can be waitten as V ant Be = 5 5 5% 2 at H, =F sink a) VX Hy = ing Fy os Es Gre. 27) RADIAL WAVEGUIDES 93 on omission of the distinguishing mode indices. ‘The characteristic impedance and mode constant of the dominant H-Lype mode are dar k AY 7 ft wt eeye-( ao matched nondissipative radial line is z Pap tt Tor (70) b. Cylindrical Sector Cross Sections.—Another example of a radial waveguide is provided by the nonuniform region illustrated in Fig, 216. (a) General view @) Site view ro, 210.—Radial waveguide of sectoral cross section. In the réz coordinate system indicated therein the cross-sectional surfaces are cylindrical sectors of aperture & and height b. E-type modes.—In the above type of radial waveguide the field ‘components of an B-type mode are EB, = —ViF cos FF 2 sin By = VE TE sin Fs cos ™ ¢, (77a) 12m ne me — a0, 3 BF sin a sin H. = 0, 12 in ™™ Hy = 1, 5 cos 3 asin B 6, 7) Hy = ~jaVi whereo TRANSMISSION-LINE MODES (Sno. 27 ‘The « dependence of this Enw-type mode is determined by the trans- mnissiow-line behavior of the mode voltage Vand current Jj, as given by Eqs. (1-64) with 7s) «79) and as before is indicative of the regions of propagation and nonpropaga- tion. "The dominant H-typo mode in the radial waveguide of Fig. 216 is the m= 1, n= mode, ‘The nonvanishing field components of this ‘mode are (omitting mode indices) V sin E B= —Fsing I yt Hy = 7 sin 5 & (60) VX set H, = ~ in| pO HS ‘th chaasterite impedance std med constant ace Be wlecley, 6 zarBh and vege =(): » od the trastisiontinebeherow ie desorbed in terms of the standing Jealkr) and Nealkr). (82) tn terme of the dominanteede rms veltage ¥ the total outward ower Ira mntched ordipative guides 2 2, aa Mor (83) “H-type moies—The field components of an H-type mode in the radial waveguide of Fig, 216 are Seo. 27] RADIAL WAVEGUIDES 95 (840) (sb) where m=0,1,2,3, 557, 1 =1,2,3, ‘The edependent mode voltage Vi’ and current 1,’ satisfy the radial trans- mission-line equations (1-64) with ay (85) ‘he eto wavelength sth aoa that ge In Ba (0) or 8 ‘Soe'molo the tl stead mode pore ben Re (PP, Tae dig ype moe te (w= w= ee nosreistng fll compan ef he Jinan ee me ar (onic ting mode indices) vs (ome vat Kanbs, a + sin % (86) Hy = win EP cost rok ‘n\* perBE mt enJe@ on ‘The transminsioning behavior ofthe dominant H-type mode i expremes in terms of the standing waves * Prose Jel) and No(ur).96 TRANSMISSION-LINE MODES [Sn0.28 total outward power flow transported by this dominant mode in a matohed nondissipsive sectoral guide can be expressed in terms of the rms current J as rok 2 2b x xmhiler) 28. Spherical Waveguides. a. Fields in Free Space—On introdue- tion of a r$ spherical coordinate system, as shown in Fig. 217, it is evident that free space may be regarded as a nonuniform trans- mission region or spheries! wave- guide. The transmission direc tion is along the radius rand the -Y cross sections transverse thereto are complete spherical surfaces desoribed by the coordinates @and #. In practice many spherical cavities may be conveniently re~ % garded as terminated spherical io, 247,—Spherical cowrdinate aystem for guides, naves nace ‘The e’-mode functions charsc- teristic of the H-medes in a spherical guide are derivable from scalar functions P 5 Io. (ss) B= We = Ff Pr lcos 6) SO (so) where de n(n tt) (tm! Ma ol (ml me 07 man neL2aoo mal itm=o, La 2 ifm, and P9(e0s 8) is the associated Logendre function of order n and degree rm. ‘Typieal Logendre functions of argument cos # are P= sing, P} = —3sin@cos 8, PE=8sin’#,? (90) P= H(5 cos? @ — 8608 @), Pi = —# sin 465 cos* # ~ 1), Ph= isin? @cos@, P} = —18 sin’ 6 Of. A. Sohelkunoff, Fletromagnetic Waves, Chaps, 10 and 11, Van Nostrand, 1048 Sec. 28) SPHERICAL WAVEGUIDES 97 By Eqs. (1:3), (1-10), and (89) the field components of the Eawmode are Vj 1 dbs (cos 8) cos Bs TM do sin ™* By = Th pr(oos 0) 8 mg, (10) . 7 Nisin @”" cos ”) EG monn me cary Palcos 0) Lk 1 dP2(cos 8) cos ~7M dd sin Hee (918) H,=0, md, where the degeneracy of the Fy,-mode is indicated by the two possible polarizations in g. ‘The r dependence of the fields is determined by the mode voltage Vjand current I{. ‘These quantities obey the spherical transmission-line equations (1-94). The components of the mode functions ef normalized according to Eqs. (1-5) with dS = sine do dé are obtained from Hy and B, of Eq. (91a) on omission of the amplitude factor Vi/r. ‘The oubward power earried by the ith mode is Re (Vslf) ‘The dominant mode in a spherical waveguide is the electric dipole mode n= 1. For the case of circular symmetry (m= 0) the non- vanishing components of this mode are (92) ‘on omission of the mode designations. In terms of the rms voltage V the total power carried by an outward traveling dominant mode is (¢f See. 1'8) ror. (93) oy OF ‘The e/’-mode functions characteristic of the H-modes in a spherical waveguide are derivable from the scalar functions shown in Eq. (89). ‘The field components of the H,mode follow from Wqs. (1-4), (1-10), and (89) as98 TRANSMISSION-LINE MODES (See, 28 aU om sin FE cain 9 Patoos BE me, ) ne, (4a) 6, ony (948) Vi n(n +1) 7 EN: Pr(co 8) 2 mo ‘The + dependence of the mode fields is determined by the spherical ‘transmission-line behavior of the mode voltage V7’ and current I”. ‘The components of the normalized mode function are obtained from Hy and E, of Eq, (948) on omission of the amplitude factor V/'/r. Lonetina view (ross sectional view Tia, 218—Conieal waveguide. ‘As for the case of modes in a radial waveguide, the concepts of cutoff wavelength and guide wavelength lose their customary significance in a spherical guide because of the lack of spatial periodicity along the trans- mission direction. ‘The cutoff wavelength 2er Vin #1) of both the Bay and Hyw-modes is, however, indicative of the regions wherein these modes are propagating or nonpropagating. For regions such that X <2, the mode fields decay spatially like 1/r and hence may be termed “propagating”; conversely for > du the mode fields decay faster than 1/r and may, therefore, be termed “‘nonpropagating.”” 6. Conical Waverwides.—A typical conical waveguide together with its associated spherical coordinate system 764 is illustrated in Fig. 218. (95) See. 28) SPHERICAL WAVEGUIDES 99 ‘The transmission direction is along the radius r, and the cross sections transverse thereto are 0% spherical surfaces bounded by cones of aperture 4, and 6, ‘The conical waveguide is seen to bear the same relation to a spherical waveguide that a coaxial guide bears to a circular guide. ‘Examples of conical guides are provided by tapered sections in coaxial guide, conical antennas, ete, ‘The dominant B-mode in the conical guide of Fig. 2-18 is a transverse electromagnetic mode whose nonvanishing components are v 1 u ‘cot (0,72) sin cot (02/2) (96) ru He Seino ‘The r dependence of the dominant mode voltage V and current I is determined by the spherical transmission-line equations (1-94); for this case of the dominant (n = 0) motie these reduce to uniform transmission- Tine equations. ‘The choice of normalization is such that the character- istic impedance and propagation wave number are aE gq 908 (02) Br cot (02/2) ond ek on ‘The total outward dominant-mode power flow in a matched nondissipa- tive conical guide is correspondingly ot (64/2) Pr ot (64/2) In Wor. (98) ‘The cutoff wavelength of the dominant mode is infinite, ‘The attenuation constant of the dominant mode in a conical guide is a function of r and is given by ft ft 1 Ss Fin on wat OY on ™ cot (02/2) where ®; and ®z are the characteristic resistances (ef. Table 1-2) of the inner and outer metallic cones. ‘The ef mode functions characteristic of the H-modes in a conical guide are derivable by Eqs. (1-3) from scalar functions of the form [P(c0s #Pz(— cos &s) — Pz(— cos 6) Pr(cos ,)] °° md, (100)100 TRANSMISSION-LINE MODES (See. 28 where the indices n (nonintegral in general) are determined by the roots of, Pr(cos 0)FZ(— c0s 04) — P3(— cos 62)P3(cos 6:) = 0, whee m = 0,123," ++ The mode faneians cartes o the H-modes are derivable by Eqs. (1.4) from scalar functions Pa(— cn sien rien) ES — pa on 9 PECAN, yon shes the nts wae the rots of Peles 6) dP on 0) Pa ~ 0 8) Pa.con 6) do co do de =o whee m= 0,2. <6. The ack of equate tabulations bath of the roots m and of the fractional order Legendre functions does not justify detailed representation of the field components of the higher modes. The special case, & = x ~ @, of a conical antenna has been investigated in some detail by Schelkunoft (le. cf.) CHAPTER 3 MICROWAVE NETWORKS 3, Representation of Waveguide Discontinuities—Waveguide struc- ‘tures are composite regions containing not only uniform or nonuniform ‘waveguide regions but also discontinuity regions. ‘The latter are regions wherein there exist discontinuities in cross-sectional shape; these discontinuities may océur within or at the junction of waveguide regions. ‘As indicated in the preceding chapters the fields within each of the wave- uide regions are usually completely deseribed by only a single propa- gating mode. In contrast the complete description of the fields within a discontinuity region generally requires, in addition to the dominant propagating mode, an infinity of nonpropagating modes. Since ‘waveguide region can be represented by a single transmission line approprinte to the propagating mode, it might be expected that the representation of the discontinuity regions would require an infinity of transmission lines. ‘This expectation is essentially correct but unnecessar~ ily complicated. The nonpropageting nature of the higher-mode ‘transmission lines restriots the complication in field description to the immediate vicinity of the discontinuity. Hence, the discontinuity fields can be elfeotively regarded as “lumped.” ‘The effect of these Tumped discontinuities is to introduce corresponding discontinuities into ‘the otherwise continuous spatial variation of the dominant-mode voltage ‘and current on the transmission lines representative of the propagating modes in the overall microwave structure. Such voltage-current dis- continuities ean be represented by means of lumped-constant equivalent circuits. The equivalent circuits representative of the discontinuities together with the transmission lines representative of the associated wave- guides comprise a microwave network that serves to describe the fields almost everywhere within a general waveguide structure, ‘The present chapter is principally concerned with the general nature and properties of the parameters that characterize such microwave networks, The determination of the fields within a waveguide structure is primarily an electromagnetic-boundary-value problem. An clectro- ‘magnetic-boundary-value problem involves the determination of the lectric fetd E and magnetic field H at every point within a closed region of space, These fields are required to satisfy the Maxwell field equations and to assume prescribed values on the boundary surface enclosing the 101loz MICROWAVE NETWORKS (See. 34 given region. According to a fundamental theorem a unique solution to this problem exists if the tangential component of either the electric field or the magnetic field is specified at the boundary surface. The reformu- lation of this field problem in terms of conventional nebwork concepts will be illustrated for # general type of waveguide structure, 1a, Impedance Pepresentalion—A typical waveguide structure is depicted in Fig. 3-1. ‘The over-all structure is composed of a discon- tinuity, or junction, region J’ and a number of arbitrary waveguide regions 1... ,N, The boundary conditions appropriate to this structure are AB T s s s , \ ‘f e Fie, 341.—Juncton of N waveguides, that the electrio-ield componente tangential to the metallic boundary surface S, indicated by solid lines, vanish and that the magnetie-field com- ponents tangential to the “terminal,” or boundary, surfaces Ts, . . . , Tx, indicated by dashed lines, assume prescribed but arbitrary values. Tt is, further assumed that the dimensions and frequency of excitation are such that only a single mode ean be propagated in each of the waveguide rogions although this is not @ necessary restriction. It is thereby implied that the terminal surfaces 7, . . . , T'y are so far removed from the junction region J that the fields at each terminal surface are of dominant-mode typs. Consequently the tangential electric fiold E, and magnetic field H, at any terminal surface 1’, may be completely charac- terized by the equations (cf. Sec. 1-3) Ex(e yin) = Vn Om Hi(ryjtn) = TnI where en and hg are the vector mode functions indicative of the cross- sectional form of the dominant mode in the mth guide, where zm denotes the outward unit vector along the axes of the mth guide, and where the bay = Zon X Cay @ Sec. 31] REPRESENTATION OF WAVEGUIDE DISCONTINUITIES 103 voltage Vm and current I, denote the rms amplitudes of the respective fields at Tr; the normalization is such that Re (Vql%) represents the average power flowing in the mth guide toward the junction region. ‘The above-quoted uniqueness theorem states that the electric field within the space enclosed by the terminal surfaces is uniquely determined by the tangential magnetic fields or, equivalently, by the currents J,, +++) Ty at the terminals 7, ; Tr. In particular the tangential lectric fields or, equivalently, the voltages Vs, . . . , Vw at the terminal surfaces Ti, .... , Ty aredetermined by thecurrents i, ... jx. The linear nature of the field equations makes it possible to deduce the form of the relations between the voltages and currents at the various terminals without the necessity of solving the field equations. By linearity it is evident that the voltages Vs, . . . , Vy seb up by the eurrent Jy, or Is, or Iy acting alone must be of the form Zuly Vie uly Via Zuvlw Zul, Vix Zl, Va = Zane, @ Zul Vx = Zeal Vw Lata) where the Zna are proportionality factors, or impedance coefficients, indicative of the voltage set up at the terminal 7, by a unit current acting only at the terminal 7. By superposition the voltages resulting from the simultaneous action of all the currents are given by Va Zul + Zale + + +Zunls, Vam Zails + Zale + > + Lawl, @) Vy = Zila + Zale $00 $2 an ‘These so-called network equations, which completely deseribe thebehavior of the propagating modes in the given microwave structure, are frequently characterized simply by the array of impedance coefficients Zar Zax +> + Za Bay Tas > Taw Le fii ee called the impedance matrix of the structure. The foregoing analysis of an NV terminal pair microwave structure is304, MICROWAVE NETWORKS (S80. 34 the exact analogue of the familiar Kirchhoff mesh analysis of an n ter- ‘minal pair low-frequency electrical structure. As in the latter ease, many properties of the impedance cocficients Zne may be deduced from general considerations without the necessity of solving any field equations. ‘The more importantof these properties are Zn = Zum anid Re (Zn) = 0 With appropriate voltage-current definitions (¢f. Chap. 2) the former of these relations are gonerally valid, whereas the latter pertain only to nondissipativestrucsures. In addition to the above, many useful proper- ties may be derived if certain geometrical symmetries exist in a wave- guide structure. Such symmetries impose definite relations among the network parameters Zs (of. Seo. 32)—relations, it is to be stressed, that can be ascertained without the necessity of solving any field equations ‘These relations reduce the number of unknown parameters and often yield important quslitative information about the properties of micro- wave structures. "The form of the network equations (3) together with the reciprocity relations Zan = Zanimply the existence of a lumped-constant equivalent cirouit which provices both a schematic representation and a structural equivalent of the relations’ between the voltages and currents at the terminals of the given microwave structure, ‘This equivalent circuit, or network representaiion, provides no information not contained in the original network eqvations, but nevertheless serves the purpose of casting the results of field caloulations in a conventional engineering mold from which information can be derived by standard engineering calculations In view of this representation, the boundary-value problem of the determination of the relations between the far transverse electric and magnetic fields on the terminal surfaces is seen to be reformulated as a network problem of the determination of the impedance parameters Znw. ‘These parameters may be determined either theoretically from the ficld equations or experimentally by standing-wave measurements on thestructure, In either case itis evident that the impedance parameters, provide a rigorous description of the dominant modes at the terminal surfaces and hence of the electromagnetic fields almost everywhere. This “far” description, of course, does not include a detailed analysis of the fields in the immediate vicinity of the discontinuities. Tn the reformulation of the field description as a network problem the choice of terminal planes is seen to be somewhat arbitrary. This arbitrariness implies the existence of a variety of equivalent networks for the representation ef a waveguide structure. Any one of these networks completely characterizes the far field behavior. No general criterion exists to determine which of the equivalent networks is most appro- priate. ‘This ambiguous situation does not prevail for the ease of lumped low-frequency networks, because there is generally no ambiguity in the Sve. $1] REPRESENTATION OF WAVEGUIDE DISCONTINUITIES 105 choice of terminals of alumped cireuit. However, even at low frequencies there are, in general, many circuits equivalent to any given one, but usually there is a “natural” one distinguished by having a minimum number of impedance elements of simple frequency variation. It is doubtful whether a corresponding “natural” network exists, in general, for any given waveguide structure. In special eases, however, the same criteria of a minimum number of network parameters, simple frequency dependence, ete., ean be employed to determine the best. network repre- sentation, ‘These determinations are facilitated by the ability to trans- form from a representation at one set of terminal planes to that at another (Ef. Sec. 33). ‘Various definitions of voltage and current may be employed as meas ures of the transverse fields in waveguide regions. The arbitrariness in definition introduces an additional source of flexibility in the network representation of waveguide structures. For example, if the voltages Vx ‘and currents Z, employed in Eqs. (3) are transformed into a new set P. and I, by I= In Vn (4a) the transformed network equations retain the same form as Eqs. (8) pro vided the transformed impedance elements are given by Bow = Zinn Vo a) ‘The new representation may possess features of simplicity not contained in the original representation. Because of this itis frequently desirable to forsake the more conventional definitions of voltage and current, in order to secure a simplicity of circuit representation. It should be noted ‘that the new definitions are equivalent to a change in the characteristic {impedances of the terminal waveguides or, alternatively, ton introduction of ideal transformers at the various terminals. b, Admittance Representation. —Although the preceding reformulation of the “far” field desoription of the microwave structure of Fig. 3-1 has been carried through an impedance basis, an equivalent reformulation on an admittance basis is possible. In the latter case the original bound~ ary value problem is specified by indication of the transverse components of the electric rather than the magnetic field on the terminal surfaces Ty, ..., Tr The introduction of voltages V», and currents Tq on the terminal planes together with a Kirchhoff analysis on a node basis (i.e. Vn rather than [y specified at 7'n) leads in this case to network equations of the form106 MICROWAVE NETWORKS (Sno, 34 T= Yue + Yue bes + Yuen, e= Yai + YaVa t+ + Yee, : . © Ty = YnVa t YouVat oo + YueVy where the admittexce elements Yaw possess the same general properties Yan = Yam and in the nondissipative case Re (Yan) = 0-88 the impedance elements of Eqs. (3. In this case the admittance element Yana represents the current set up st the terminal 7’ by a unit voltage applied only at the terminal T. As an alternative to Eqs. (5) the array of admittance coefficients Yu Yass + Yow Ya Yass >> Yew vel a You Yous ++ Yea! called the admittance matrix of the waveguide structure, is sometimes employed to charueterize the dominant-mode behavior of the given structure. ‘The statements relative to the arbitrariness in choice of terminal planes and veltagecurrent definitions apply equally well to the admit- tance description. However, the equivalent network representation of the network equations (5) is dual rather than identical with the network representation of Kas. (3). . Scattering Representation.—An alternative description of the fields within the waveguide structure of Fig, 3:1 stems from a reformulation of the associated felc problem as a scattering problem. Accordingly, in addition to the general requirement of the vanishing of the electrio-eld components tangertial to the metallic surfaces, the original boundary- value problem is defined by specification of the amplitudes of the waves incident on the terminal planes T, .. . , Ty. In this scattering type of description the dominant-mode fields st any point in the waveguide regions are described by the amplitudes of the incident and reflected (scattered) waves at that point. In particular the feds at the terminal plane 7’ are described by E:Guitn) Hiys80) he = Fon X Go } ° where ey, and hy are vector mode functions characteristic of the trans- verse form of the dominant meda in the mth mide and a nnd ham Sue. 31] REPRESENTATION OF WAVEGUIDE DISCONTINUITIES 107 respectively, the complex amplitudes of the electric field in the incident and reflected wave components of the dominant mode field at T. The normalization of the mode functions is such that the total inward power flow at any terminal 7. is given by Jan|* — |bq|?; this corresponds to a choice of unity for the characteristic impedance of the mth guide. ‘The fundamental existence theorem applicable to the scattering formulation of a field problem states that the amplitudes of the scattered ‘waves at the various terminals are uniquely related to the amplitudes of the ineident waves thereon. As in the previous representations the form of this relation is readily found by adduetion of the linear nature of the clectromagnetic field. Because of linearity the amplitudes of the refiected waves set up at the terminal planes Ts, . . . , Ty by a single incident wave a; at 7', or zat Ts, . . . are b= Sua, = Sud, br = Svar, br= Sum, by = Siete, br = Sern, a) by = Sma, by = Suds, by = Swat. Therefore, by superposition the amplitudes of the scattered waves aris- ing from the simultaneous incidence of waves of amplitudes a, .. . , ay are by = Sas + Stas + + = + Swe, bs = Saas + Sua + + + Sivas, oe : Y ® be = Siar + Seats + + + Sunde ‘where the proportionality factor, or scattering coefficient, Sais a measure of the amplitude of the wave scattered into the mth guide by an incident ‘wave of unit amplitude in the nth guide. In particular, therefore, the coefficient Snm represents the reflection coefficient at the terminal 7 when all other terminals are “matched.” For brevity it is frequently desirable to characterize the sttering properties of a waveguide structure by the array of coefficients Su Sin + Sw Sx Sea Sie Sef oliil id ® \Sv1 Sx2 ++ + Sy soled the eonttorine matete oothor Ahan toe Moe £2108 MICROWAVE NETWORKS [re 92 ‘The clements Sau of tho scattering matrix S may be determined either theoretically or experimentally. ‘The values so obtained are dependent on the choice of terminal planes and the definitions of incident and ‘scattered amplitudes. Certain general properties of scattering coeffi- cients may be deduced from general considerations. For example, with the above definitions [Eqs. (6) of the amplitudes a, and b,, it ean be shown that 1. The reciprocity relations Spa Sony x 2. The unitary relations ) S*mnSs= (20) A are valid; the latter apply only to nondissipative structures, If the given structure possesses geometrical symmetries, itis possible to derive corresponding symmetry relations among the scattering coeffi- cients. These relaticns, derivable without the necessity of solving field equations or perform-ng measurements, are identical with those for the elements of the impecance or admittance matrices of the same structure. ‘The reformulation of fcld problems either as network problems or as scattering problems provides fully equivalent and equally rigorous deseriptions of the far field in a microwave structure. ‘The choice of ‘one or the other type of description is difficult. to decide in many eases. In favor of the impedance or admittance descriptions are the following, facts: (1) The descriptions are in close accord with conventional lov~ frequeney network descriptions; (2) they can be schematically represented by equivalent circuits; (3) they lead to simple representations of many series or shunt combinations of discontinuities and junctions. In favor of the scattering description aro the facts: (1) It is particularly simple and intuitive when applied to the important case of matched or nearly matched microwave structures; (2) reference-plane transformations ean. be effected quite simply by phase shifts of the scattering coefficients. For the most part impedance or admittance descriptions are employed ‘throughout the present volume sinee itis desired to stress the connection ‘between microwave network analysis and the conventional low-frequency network analysis. For interrelations among the various descriptions the readeris referred to Prineiples of Microwave Circuits by C. G. Montgomery and R. Dieke, Vol. 8 of this series. 32, Equivalent Cxcuits for Waveguide Discontinuities—The Kirch- hoff analysis of the far fields within a general N-terminal pair microwave structure can be expressed in terms of N(N + 1)/2 complex parameters [ef. Eqs. (3) and (5)] and represented by a general N-terminal-pair equivalent network, If the structure possesses geometrical symmetries, itt is possible to reduce the number of unknown network parameters and Spe, 32] EQUIVALENT crRcurTs 109 correspondingly simply the form of the equivalent network by means of a Kirchhoff analysis that utilizes these symmetries. Symmetrical struc- tures are characterized by the existence af two or more terminal planes looking into any one of which the structure appears electrically identical. AAs outlined in the preceding section, a Kirchhoff analysis of the response due to current excitation at one of these symmetrical terminal planes is described by impedance coefficients given by one of the columns in Eqs. @), The columns describing the responses due to current excitation at the other symmetrical terminal planes can be expressed in terms of these same impedance coefficients, but in different order. It is thus evident that the symmetry properties of the given structure can serve to reduce the number of unknown impedance coefficients. Results of analyses utilising structural symmetries will be tabulated in this section for several microwave discontinuities. The Kirchhoff analysis of a symmetrical microwave structure can be effected on either an impedance or an admittance basis. ‘The choice of analysis is generally dictated (at least for the A’ > 3 terminal pair structures) by the type of geometrical symmetry possessed by the struc- ture. Tt is not implied hereby that only one type of description is possible in a given case. An impedance or an admittance description is always possible, In a structure with a certain type of symmetry the ‘impedance description, for example, may be found most desirable since the parameters of the equivalent circuit for the structure may be simply related to the elements of the impedance matrix but not to those of the admittance matrix. ‘The possible existence of another equivalent circuit ‘whose parameters are simply related to the elements of the admittance rather than of the impedance matrix is not excluded. However, the ‘two equivalent circuits will not, in general, be equally simple. ‘The preferred description is that based on the simplest equivalent circuit. In the following the equivalent circuits together with the correspond- ing impedance, or admittance, representations of several general classes of microwave structures will be presented. No detailed effort will be made either to show how the symmetries of the structure delimit the form of the matrix and circuit representations or to discuss the reasons for the choice of a particular representation. ‘The consistency of a representation with the symmetry of a structure ean be readily verified on application of e Kirchhoff analysis both to the given structure and to ‘the equivalent circuit. These analyses lead, of course, to the same ‘matrix representation. Incidentally the recognition of the applicability of conventional Kirchhoff analyses to microwave structures constitutes ‘an important engineering asset, for one can thereby set up and delimit the impedance or admitianco matrix or, alternatively, the equivalent cirouit representation thereof and derive much information about the110 MICROWAVE NETWORKS See. 82 behavior of a given microwave structure without the necessity of solving. any field equations. For a given tye of geometrical symmetry the equivalent circuit information will be seen to apply equally well to a variety of waveguide structures of which only a few will be pointed out. Since no specific choiee either of terminal planes or of voltage-current definitions will be made, the representations to be presented are of a quite general form and can be considerably simplified by a judicious choice of these factors (cf. Sec. 3:3). However, the positive directions of voltage and current will be indicated since the form of the impedance or admittance matrix ee o @ A Cee 8) Symmetrical junction of two guides a) Syrmatial scantily ina waveguide h Me be Ya Me Ys Ta q % (@Aditarce mat (dpequvalet crest Fie, 3.2 Symmetrealtwo-terminal-pir waveguide strcture. (though not the equivalent circuit) depends on this choice. Furthermore ‘the location of the terminal planes, though arbitrary, must be in accord. with the symmetry of the given structure, It is assumed throughout that only the domiaant mode ean be propagated in each of the wave- guides, this unnecessary restriction being employed only for the sake of simplicity a, Two-terminal-pair Networks—Typieal two-terminal-pair wave- guide structures of arbitrary cross section are illustrated in Fig. 3-20 and b. For a symmetrical eaoice of the terminals 7, and 7 relative to the central plane, the symmetry of the structure imposes a corresponding symmetry on the admittance matrix and equivalent circuit representation of the over-all structure. ‘The general representation of a two-terminal-pair structure is thereby reduced to thet shown in Fig. 32 and d. The positive directions of voltage and current have been so chosen as to ‘obtain positive off-ciagonal elements in the admittance matrix. See. 32] RQUIVALENT crecurTs un ‘The simplification in circuit description resulting from the symmetry of the above two-terminal-pair structures can be taken into account equally well on an impedance basis. In this ease the circuit represen- tation is expressed in terms of a symmetrical T-cireuit rather than of the s-circuit employed in the admittance description. b. Three-terminal-pair Networks—An arbitrary junction of three waveguides may be represented by either the impedance or the admi tance matrix shown in Fig. 3-3aor5. The equivalent circuits correspond- fy es wy te Me Za tn Ze Ye Tele Bs ta ty hee (ioc mati gel arinitanc mates!» SEC at ere er T Otttatet 34 Tia. 33.—Representation of genera three-terminal-pair structure ing to these matrices are dual to one another and can be represented as indicated in Fig. 3c and d. ‘These representations can be considerably simplified for the case of symmetrical structures ‘An important class of symmetrical three-terminal-pair structures is that in which geometrical symmetry exists with respect to. plane. Such symmetry implies that the symmeizy plane bisects one of the guides, the so-called stub guide, and is centrally disposed relative to the remaining two guides, the latter being designated as main guides. Structures with this planar symmetry may possess cither E- or H-plane symmetry, depending on the type and relative orientation of the propagating modes in the main and stub guides. E-plane symmetry obtains when sym- metrical lectric-eld excitation in the main guides results in no coupling to the stub guide. On the other hand, H-plane symmetry implies that112 AUCROWAVE NETWORKS [Se0. 32 antisymmetrical cleotrie-feld excitation in the main guides produces no excitation of the stub guide. It should be noted that such properties are rot present if modes other than the dominant ean be propagated in the stub guide. STRUCTURES WII E-PLANE syunceray.—Two junetions with E-plane symmetry are illustrated and represented in Fig. 3:4. When formed of guides with rectangular eross section, such junctions are characterized St AK (a) Syemetical E-plane Function (0) SymmeticlE-plane Function —y% ( Admitance matic (a Eauttant cect Fro, 84-—Symmtrcalthre-terminal-pairstrusturee—Replana eymmetcy. by the fact that the far electric ficld is everywhere parallel to the plane of the above figures. The indicated admittance matrix and equivalent circuit representations of such structures depend on a symmetrical choice of terminal planes in guides (1) and (2). For the special ease of ¥ junotion with © = 120°, it follows from the added symmetry that Yu = Yi; and Yir = Yas, provided the terminal plane in guide (3) isselected in the same symmetrical manner as those in guides (1) and (2). ‘The equivalent circuit of Fig. 3-4d therefore reduces to that shown in Fig. 3-ba. For the case of a ¥ junction with ® = 0, theso called E-plane bifurea- tion, the sum of the terminal voltages is zero and consequently Yy = = G, J = 1, 2, or 8) if the terminal planes are all chosen at the plane of the bifurcation. Although the admittance matrix is singular in this case, See. 32] BQUIVALENT CIRCUITS 43. differences of the matrix elements are finite and the equivalent eireuit of Fig. 8-4d reduces to that shown in Fig. 3:55. For a bifurcation with a Cy ® Fro. $5-—(@) Equivalent ereult for 120% Y-junetion. @) Equivalent aeuit for 0° ‘junction. dividing wall of arbitrary thickness ¥ = Yet Yor — 2% ie, and for a wall of nero thickness _Yu>¥u = STRUCTURES WITH T-PLANE sysacerRy.—The sectional views of the junctions illustrated in Fig. 3:4a and b apply as well to junctions with H-plane symmetry. In the latter case, for guides of rectangular cross section, the far magnetic field is everywhere parallel to the plane of the figures. Coaxial T and Y junctions, though not possessing the same geometrical structure, have the same type of field symmetry; junctions of this type are illustrated in Fig. 3-6 and b. ‘The associated imped- ‘ance matrix and equivalent circuit representations shown in Fig. 3:6 correspond to symmetrical choice of terminal planes in guides (1) and (2). If & = 120° in tho H-plane ¥ junctions of Fig. 8-4, the higher degree of symmetry implies that Zz = Zia and Zi: = Zz, provided the terminal plane in guide (8) is chosen symmetrically with those in guides (1) and (2). y=4 MICROWAVE NETWORKS [Sr0, 32 In this case the equivalent network of Fig. 36d becomes completely symmetrical and is composed of a common shunt arm of impedance Zra and identicel series arms of impedance Zin — Zs. _t __ _ LL a 2 Qj a mee meee © consi Bnei (a) Probe coupled junction ofcsaxal ahd arbivary wavezides (© Impatance matic {4 Fault cic ® Fro, $:6—Symmesicl three terial pie steueturesH plane eymmetry. , Fourserminal-pair Networks, JUNCTIONS WITH R-PLANE. SYM acermy.—Junetions o! four rectangular guides with B-plane symmetry are indicated in Fig. 37a and b. Since there exist two symmetry planes, cither guides (1) ané (2) or guides (8) and (4) ean be designated as the rain guides or as thestub guides. The designation B-plane is consistent with the fact that tae far electrie-field intensity is everywhere parallel to the sectional plane indicated in the figure. If guides (1), (2) and azides (8), (4) are identical and the terminal planes in identical guides are chosen symmetrically, the admittance matrix and equivalent circuit representations of tho structure are shown in Fig. 376 and d. ‘The indicated equivalent circuit applies to the junction in Fig. 87% only if the thickness of the dividing wall is sufficiently large to make negligible the B-mode coupling (.c., the coupling resulting when the normal electri field is @ maximum at the aperture). If this situation does not prevail, as is the ease when the thickness of the dividing wal is small, the +45" diagonal elements of the admittance matrix of Fig. 8-7e should be changed from Yis to Ys in order to take account of both E- and H-modes of coupling through the aperture. For the case of four identical guides, a dividing wall of sero thickness, and all terminal planes See. 32] EQUIVALENT crecurTs 15 =, af f 1 a = 2) 1 ‘ of ! @ 7 0 pnt or ® “anton (Ene ot icon ny tor nn Ya Ya Ta Ms Ke Ys Ta Ye Yaa t—w—f Yaa Ya Ya Ye Ye oy — (0 Aémitance mata ‘Eqvatent chet io, $7.—Symmetrical four-teminal-peireteustures—it-plane symmetry. r Te | Ye % chosen coincident at the central reference plane 1’, the elements of the admittance matrix become infinite but differ from one another by a finite amount. The equivalent cireuit corresponding to the resulting singular ‘matrix is illustrated in Fig. 3-8 where Yo = 2(¥n — Yu) = 2(¥ua ~ Yoo) = 2(¥os — Fo), ¥y = 20 ~ Yu) = 2% ~ Yu),116 MICROWAVE NETWORKS (Seo, 82 JUNCHIONS WITH H-PLANE synurRY.—The sectional views of Fi 3:7a and b apply equally well to junctions with H-plane symmetry; such junctions the far magnetic field intensity is everywhere parallel (ie, the far electric fied is everywhere perpendicular) to the plane of the indieated sectional view. The coaxial guide junctions shown in Fig. 39a and b also possess the same field symmetry as the H-plano junctions. o Oy {a @ ra Bo B EE eH He PAHO (8) cross veto of two coal ules {a Cont waves junction Zn te ty Ba 2a te Bn Zia faa Tay ae Zn fy Zu 20 (impedance matic (@exunatetcreut Fre, $9:—Syimmetralfourterminal-pair sructures— Hl lane symmetry. If the terminal planes 1 and 7 are chosen symmetrically (as likewise Ts and T,), the equivalent cirouit and impedance matrix representations of this class of structures are shown in Fig. 39 and d. If the dividing wall in the H-plane junction of Fig. 37b is of small thickness, the +45° diagonal elements of the impedance matrix are to be changed from Zs to Zw. ‘The corresponding equivalent circuit is shown in Fig. 3-102, The special case of identical guides, a dividing wall of zero thickness, and all terminal planes coinciding at the central reference plane ‘Tis represented by the equivalent, cireuit of Fig. 3-106. Magic TsUNcTIONS-—Two typical magic T-junctions are depicted in Fig. 31a and 6. In Big. $'Lla a symmetrical junction of four rectangu- lar guides is illustrated in which guide (8) is the H-plane stub and guide (4) is the E-plano stub. Figure 3:11b is a symmetrical junction of one See. 38} EQUIVALENT REPRESENTATIONS 47 AZ) -(Eae2d Ba) Coe Z) AyeNga Zo) MeymZy) () att atten sree (0) wat of ro tia Fro, 310. Equivalent circuits for junction of Fig. 872-—H-lane symmetey. Ban 2-8, 2) i" a) fy uO tw (2 tmpeanc mate (a eaten est Fee, $41.—Magie 7 structures coaxial and three rectangular guides; in this figure the coaxial guide (3) is the H-plane stub, If the terminal planes in the identical guides (1) and (2) are chosen symmetzieally, the impedance matrix and equivalent, ixeuit representations of these junctions are given in Figs. 3-1ic and d. 3-3, Equivalent Representations of Microwave Networks.—Many of the equivalent circuits indicated in the preceding sections may be unsuit- able‘in practice either because of difficulties in carrying out networkus MICROWAVE NETWORKS [Bre. 33 computations or bectuse of the complexity in the measurement and frequency dependene2 of the eireuit parameters. By an appropriate choice both of voltage-current definitions and of reference planes, alterna- tive circuits can be devised in which such difficulties are minimized. Several equivalent representations obtained in this manner will be described in the present section. Since symmetrical N-terminal-pair representations ean often be reduced by symmetry analyses (bisection Seo. 33] EQUIVALENT REPRESENTATIONS 119 where WL Yuden = Yhap Ml = Sanden ~ Zia = oy "The relation between the ioput impedance 2 (or input admittance Yi) at Ty and the output impedance Zax (or output admittance You) at Ts is given by Yin (re-ta~ yok) an Dy 8. Ba Bix = r Es 1 At the same terminals 7’; and 1’; alternative representations of the 1 ; | H TT | above discontinuity are provided by the series-shunt cirouits of Fig. 313, oy it 1 1 1@ | ' tot 1 rt vayen hy wR OR mu TF bs 2 F g Bs 3:8: »3:6: Qu Fro, $12-—Ganeral troderminsl-pair waveguide dicontinuity. & aye de ae theorems) to & number of two-terminal-pair networks or less, equivalent * na ROR yalit 8 representations of the basic two-terminal-pair structures will be con- fn Fn Shuntanion Series shunt sidered first ‘The arbitrary disontinuity at a junction of two different guides illustrated in Fig. 342 is an example of a general two-terminal-pair microwave structure. ‘The over-all structure may be represented by transmission lines of characteristic impedances Z = 1/Y,and Zs = 1/Yx, connected at the term:nal planes 1, and 1's by either the 7 or « equivalent Bunt Zp-tn 4 a 2a ap dee B equlont % R -r eqialent te io, 8184~Cirenit mpresntations of generl two-terminal-pai structure. cireuit indicated in Fig. 3-134. ‘The relations among the circuit param- eters of the T and representations at the terminals 1; and Ts are Yu = Yur _¥_ e225 Su, fn — fae By Yu Ya om Yu= 7 a ia ay M8) Ya Yum Ss, Fra, 8188,—-Cieult representations ofa general tno-terminalpair structure, These dual circuits are equivalent to the 7 and w-cirouits shown in Fig, 3:14. ‘The primary-secondary tums ratio of the ideal transformer is denoted by 1/1; the corresponding impedance ratio at the transformer terminals is n*/1. It is evident that for a structure in which the determi- nant |2| or |¥| vanishes, the equivalent circuit becomes either purely shunt or purely series, respectively. In this special ease the ideal trans- former ean be omitted if the char- 2 a acteristic impedance of the output line is changed to n*Z2; ie,, if the voltage-current definitions in the output guide are changed. » Me ie ‘A variety of other equivalent ¢——=-———— representations for two-terminal- "9444 ~tranamissiontine roprecentar ton ofa symmetrical” twotermina-pait pair structures can be found by to employing transmission lines as circuit elements, ‘Thus, as shown in Fig. 9-14, a transmission line of length @ = sl and characteristic impedance Z» can be employed to represent & symmetric discontinuity structure with Zix = Zaa¥n = Ya). In terms of the parameters of the circuit representations of Fig. 3180 and B, the ‘ransmission-line parameters are120 MICROWAVE NETWORKS (sre. 33 t= VOT, Yon VET tan 5 = tong = foarte 2 NZ + Bw 2° V¥aF Yn ‘The consideration of the corresponding representations for asymmetrical structures will be deferred until the closely related question of the trans formation of reference planes is treated. ‘Transformations of Reference Planes—Equivalent circuit representa~ tions of a waveguide discontinuity may be considerably simpler at one set of terminal planes than at another. The investigation of simplifications of this type requires the ability to determine the equivalent circuit param- ters at one set of reference planes from the knowledge of the parameters fat any other set. For the case of the structure shown in Fig. 3:12, shifts of the input terminals from 7; to T{, a distance d away from the junction, and of the output terminals from to 7%, a distance » away From the junction, can be accomplished in several ways. A straight forward Way of effecting this shift involves the addition of transmission fines (or their equivalent circuits) of lengths d and s to the input and out- put terminals, respectively; the characteristic impedance and propagation ‘wave number of the input and output lines being Z, = 1/Y, m1 = 2r/Ayuy and Zy = 1/Yo «2 = 2r/dys, ‘The computation of the “shifted” param- ‘eters can be carried oat by standard circuit techniques. ‘Though some- ‘hat laborious, this method has the virtue of being applicable to Netorminal-pair structures involving both uniform and nonuniform {ransmission lines, Phase shift of the scattering matrix of a microwave structure provides ar alternative method of reference-plane transfor~ mations, but this will not be discussed herein, ‘For the particular sase of uniform lines there is another way of effect- ing the desired transformation, This method is based on the faet that an arbitrary two-terminal-pair network can be represented as an ideal fransformer at certain “characteristic” reference planes. Since reference plane transformations to and from these characteristic” terminals can be readily aceomplisned, @ simple means of carrying out arbitrary ‘transformations is thereby provided. ‘The existence of an ideal transformer representation of the two- terminal-pair structure of Fig, 3:12 follows from the fact that at the terminals 7 énd 7; the input-output relations of Kgs. (12) can be rewritten in terms of :hree new parameters Do, So, ¥ 98 tor m(D — De) = 7 tan ag(S — So) as) if the change of variazles Zu = ~ 52s tan xD, Dus = Hl tan ve8, Ya = 4i¥ cot nD, You = ~i¥ 2 cot «28 o 5x0. 35] EQUIVALENT REPRESENTATIONS 1a is made. ‘The relations between the parameters De, Ss, 7 of the tangent relation (13) and the parameters of the T' or x representation of Fig. 3:13, are given either as ota, Boa re) as) Leattede, iterate BF ac) a Jot where for the 1 Representation Representation tan 2 scot 25 a= tan F Dy a= ~ cot De 2 = tan 2 5, Botany, Zu an 1% an Zu c= i pa Zula = Fy pe YnYa- Yh la Ti¥s ‘Tho relations (15) are determined by expansion and identification of terms in Eqs. (12) and (13); Eqs. (16) follow from Eqs. (18) by inversion Equations (16) are not valid for the degenerate case a =f = 0, as is to be expected from the corresponding degeneracy in the impedance representation of an ideal transformer. The + signs in Eqs. (16) indicate the existence of two sets of a, 6, y equivalent to a,b, ¢; these sets are positive or negative reciprocals of each other. For each value of + given by Eqs. (16), the corresponding set of values for a and 8 may be obtained from bey, gta) +a+e% etary c-a122 MICROWAVE NETWORKS [Be0. 38 a-- pe Wt + Ot by ve — ab) For each value of a, the corresponding set of values of @ and y are baa boa “aFe’ * ata ot eb For nondissipative structures with purely reactive output impedances itis evident from Egs (14) that both D and S are real. ‘The quantity D is then the distance from the terminal to a voltage node in the input line and is counted rositive in the direction away from the junctions; —— 4 Ze 3 “ oo __] fe 2 deat vanstormee Fic, 846:—Ideal transform ropresantation of a ondiasipativetwo-terminal-paratrustre St tharatariatle relorence planes, correspondingly, S is the distance from T's to a voltage node in the output line and is also positive in the direction away from the junetion. ‘Thusif is written as —n*Zs/Z,, Eq, (13) states that “characteristic” terminals T? and TY exist, distant De and Sp away from 7; and Ts, at which the input impedance is a constant n® times the output impedance. There- fore at the terminals 7 and ? the equivalent circuit of the nondissipative waveguide junetion shown in Fig. 3:12 is the ideal transformer depicted in Fig. 3415, ‘The equivalence between the transformer representation at T?, T? and the T or x representation at T', scan be rephrased as an equivalence at the same set of terminals. For example, if lengths Do and Se of input ‘and output transmission lines are added, respectively, to the terminals 7, and T; of the T o: + representation of Figs. 9:13, a representation is ‘obtained at 13 and 72 that is equivalent to the transformer representation of Fig. 3:15. Conversely, if lengths —Dy and —S) of input and output lines are added to the terminals T?, T3 of the transformer representation, Gf, A, Weisafioch, Helfreg. w. Elektr, vol, 60, 1942, pp. 67 e sp. Seo. 334 EQUIVALENT REPRESENTATIONS 123 ‘a representation is obtained at T,, T's that is equivalent to the T or x representation. ‘The transformer representation embodied in Eq. (13) provides a relatively simple means of determining the parameters a’, D', and ¢ (cf. Eqs. (17)] of a network representation at any terminals 7%,'1°; from the corresponding parameters a, 6, and ¢ of a representation at the ter- minals 1’, T;, Let it be assumed, as indicated in Fig. 3-12, that the refer~ ence planes 74 and 7 are located at distances d and s, respectively, away. from T; and Ts. The form of the tangent relation relative to the new reference planes 7% and T{ can be readily obtained from that at the reference planes 7; and 7 by rewriting Hq. (13) as tan «{(D ~ d) ~ (Dy = d)] = 7 tan n(S — 8) — (So — 9). (18) Comparison of Eqs. (13) and (18) indicates that relative to the new terminals 7% and 7 the parameters a’, 6’, and 7’, as defined in Eqs. (17), are given by a = tan (De ~ @) BY = tan ne(So — 8) = (19) » where (20) At the new terminals 7%, 7% the relations between a’, ¥,c! and o’, 6, »/ are the same (except for the prime) as those between a, b, c and a, 8, given in Eqs. (15). The elimination of 6,» from the primed relations by means of Eqs. (19), followed by the use of His. (15), leads to the desired relations [Eqs. (21)] betwen the shifted and original network parameters. x 2 sz % Ng ane De Ape he oe dhe i % a qt @ kel ® bet Fro, 816-~Bavivalnt representations of sited wo-armialpi atwork, On transformation to new terminals 7 and 7%, located at distances dand s from 7’, and Ts, the two-terminal-pair networks indicated in Figs. 3-13a and b can be schematically represented as in either Fig. 3-16a or 3, The boxes represent networks of the 7’ or x type or any of their124 MICROWAVE NETWORKS (See. 83 equivalents; the heavy lines represent lengths of transmission lines. ‘The relations between the parameters a, b, ¢ of the original T(x) repre- sentation and the parameters a’, b,c’ of the transformed T(z) representa tion are at a0 + Bob — ante TH ava = Boe = ecfib! @) where ‘Equations (21) apply as well to the case whore a’, b,c’ are parameters of a T(x) representation and a, b, ¢ are parameters of a x(T) representation, provided the relation between ay, 8¢ and d, ¢ is It is to be noted that the two distinct sets of transformation relations distinguished by the parentheses in the preceding sentences are dual to ‘each other. ‘Asan illustration of the use of Eqd. (21) let it be required to determine the shifts d and s of the input and output terminals of the waveguide structure of Fig. 319 in order to transform the representations of Fig. a — a % % % xe whe nQ fr dhe Me de ei % Rm Es » ® © 5:17.—aulvalent representations by abit of terminal planes. (0) Shunt cepree sontaton af abieary swoscrninalpait natwork shown in Figs.3'180 and b.@) Seles ‘Spvecentation of srbisary tvoserminal-paie network shown In Pigs 130 and 0 3-13a and b into the pure shunt (series)* representation of Fig. 3-17a and. Let a, b, c be the perameters, as defined in Eqs. (17), of the original representation, and correspondingly let a’, 0’, o' be the impedance (admit tance) parameters of the transformed representation. Since for a shunt "The fllowing statements apply to the cases either within or without parentheses, respectively. Bee. 33) EQUIVALENT REPRESENTATIONS 125 (series) representation b/ = 0, it follows from Eqs. (21) that for an arbitrary ao b= aw "a a’ (22a) where = tan ao = tar xe Bo (22) for a T(e) to shunt (eres) representation, or (220 for a T(z) to series (shunt) representation. On substitution of Eq. (220) into (21) the parameters of the shunt (series) representation become ve (+ a)* + (b ~ axe)? TH ee be) Fall Ft = w= BY @ (ata)? + = ae)? oUF a) O Fac) where for the (220) Shunt Representation Series Representation @ iy omy, Both the shunt impedance jX and series admittance B of the transformed roprotentation are shown in Fig. 3-17a and 6 (also cf. Fig. 3-135). A further simplification of the transformed representation is obtained on removal of the ideal transformer by modification of the output character- istic admittance, a procedure indicated previously. ‘A useful special case of the above transformation occurs when the original network is symmetrical (a= ¢) and the input and output ‘terminals are shifted by equal amounts d = s. ‘The shift d = s required to seeure a shunt (series) representation is given by Eq. (22a) as a = Bo = a+ VOTH Therefore, Eqs. (22) become vera 1-6" 23)126 MICROWAVE NETWORKS (S20. 83 ‘These expressions assume greater significance when one notes that, Vb + represenis the shunt reactance (series susceptance) of the original symmetrical T’ (z) nebwork. ‘Another simple case obtains when d= 0. From Eas. (22) it is evident that for a T(x) to shunt (ceries) transformation (24) te a= ite, 2) ¢ite oo bFe Neterminal-pair Structures-~Equivalent representations of an N= terminal-pair waveguide structure ean be obtained either at a given set of reference planes or at shifted reference planes. In the former case repre- sentations of the type depicted in Figs. 3"13a and b can be employed to secure equivalent representations of the two-terminal-pair networks that compose the over-ll N-terminal-pair network. In addition multiwind- ing ideal transformers can be usefully employed. Since no impedance or admittance descrigtion of an ideal transformer exists, its description ‘must necessarily be phrased in terms of terminal voltage and currents. For the case of a three-winding ideal transformer, illustrated in Fig, 3"18a, the network equations are Vili + Valet Vale = ‘| (26) here n:, na, and nsare proportional to the number of turns on the various windings. ‘The relation between the input admittance ¥y at the ter- minals 7 and the output admittances ¥, and Ys at the terminals 1' and Tz follows from Bop. (26) as i See, 33] EQUIVALENT REPRESENTATIONS 127 ‘The threo-winding ideal transformer is evidently a natural generalization of the familiar two-winding ideal transformer. A network utilising the ideal transformer of Fig. 3-184 is shown in Fig. 3:18. ‘The terminals T can be regarded as terminals of an output Tine or of @ lumped-constant eirouit element. Accordingly Fig. 3-18) nob nob & ® Feo, 318—(2) Throewinding ideal transformer. (0) Network Sith theceowinding ideal represents either a four- or a three-terminal-pair network. ‘The corre- sponding network equations follow from Eqs. (26) as (Va Vols + Wat Vola t Vals = 0, Vir Vs _ Vat Vi eho (es) ‘The relation between the input admittance Ym at terminals 7's and Out- put admittances Ys, Ys, and Yat terminals 71, 7, and T¢'s given by (ry + na) VY ob nV AY ct nbVY ¥ ware + Va + Yo) (29a) ‘The input admittance Yw at 1’ in terms of output admittances Ys, Ys, and Ysat Ty, Ts, and Ts is (na + ma)*Vi¥e + iF + Yad ¥s nis + nV, + 0g¥ a Yuu (296) ‘The input admittance Yu at 7 in terms of output admittances Ys, Ys, and Y,at Ty, Ts, and Tis YY + n§(Fs + YOKs Ye Gant mT + mV FAY (290)128 MICROWAVE NETWORKS [Be0. 33 ‘The special case n= nz describes a hybrid coil, a network frequently employed to represent @ inagic T (cf. Fig. 3:11) at appropriate reference planes. For this esse it is apparent from Eqs. (29a) and (6) that when Y, = ¥,, the input admittance at 73(7') is independent of the output admittanee at T«(P). This important property forms the basis for many applications of tho magie T in bridge circuits, ete. By the use of additional elements in the network of Fig. 8:18b it is possible to obtain % no 2% 5 % 5 so © s— % % o— 2B g " @ @ Fro, $19—Steps in releenceplane transformation of representation in (e) to represen ion iO). equivalent representation of arbitrary three- or more terminal-pair networks, ‘Equivalent rep-esentations of an N-terminsl-pair structure can also be obtained by transformation of reference planes. Reference plane shifts can be effected quite simply if the over-all network repre- sentation can be redueed to a number of two- (or Jess) terminal-pair networks. Such reductions, for the case of symmetrical N-terminal- pair networks, can be accomplished by the use of symmetry analyses (bisection theorems). The reduced networks are fully equivalent to the over-all network in that the former compose the latter and conversely the latter reduce te the former. If the reduced networks are two-ter- minal-pair networks, the transformation equations (21) can be employed ‘to secure new representations at other terminal planes. With the know!- Sxe. 33] EQUIVALENT REPRESENTATIONS 129 ‘edge of the new representations of the reduced networks the over-all network ean be composed at the new reference planes. For illustration a transformation process will be employed to obtain simplified representations of the three-terminal-pair structures depicted Figs. 3-4 and 3:6, The over-all network representations a the terminals 1, Ts, and Ts are reproduced in Figs. 3-192 and 2-20a with a somewhat different notation. On bisection of the over-all networks by placement 4 % % — * 4 or 4 x eS i BOR Ri o © @ Fe, 8 20-—Bpe in eefernceslane transformation of representation ia (a) to repreea- ‘aioe nO. of a short or open circuit at the electrical centers, the reduced networks indicated in Fig. 3°19 or ¢ and Fig. 3-206 or ¢ are obtained. Simplified representations of these reduced networks can easily be found. On appropriate shift of the terminal 7) a distance d away from the junction, the reduced network of Fig. 3-19 becomes the open circuit indicated in Fig. 19d; and correspondingly, the network of Fig. 3:20 becomes the short circuit of Fig. 3:20. In addition the terminal 7 ean be shifted a distaneo s away from the junction so as to transform the reduced networks of Figs. 819% and 3-202 into the series and shunt representations of Figs. 3:19e and 3:20. ‘The over-all network representation at the new ter rminals 7%, 74, and 7%, as shown in Figs. 3-19f and 3-20, is then composed by recombination of the transformed reduced networks, ‘The relations between the parameters a, d, and ¢ of the original and a, ¥, and oof the ‘transformed networks are given by Eqs. (22a) and (c), where in this ease130 MICROWAVE NETWORKS [S0. 34 For Fig. 8-19 For Fig. 3:20 . (30) be oars oy 2s coRy eo" ‘As mentioned above the representations in Figs. 3-19 and 3-20/ can be further simplified through removal of the ideal transformer by a suitable modification of the sharacteristie impedance of the output line at ter minal 7 34. Measuremeat of Network Parameters.—Tho experimental determination of the N(N + 1)/2-network parameters that characterize an N-terminal-pair waveguide structure involves the placement of known impedances at V — | “output” terminals and measurement of the result- ing impedance at the remaining “input” terminal. A variable length of short-circuited line provides a convenient form of output impedance. A standing-wave detec:or or its equivalent provides a means for the meas- urement of input impedance (ef. Vol. 11 of this series). Input imped- ‘ance measurements must be performed for N(N + 1)/2 arbitrary but independent sets of output, terminations. The determinations of the network parameters ‘rom these measurements can be considerably simpli- fied by a judicious choice and placement of the output impedances. For example, the placement of arbitrary but fixed output impedances at N 2 terminals reiuees the overall network to a two-terminal-pair network, the parameters of which can be readily measured. A proper choice of the fixed ovtput impedances gives rise to two-terminal-pair net~ works from whose measured parameters the unknown N(A + 1)/2 param- eters are easily determined. ‘The proper choice of output impedances is generally apparent from the form of the equivalent circuit for the over~ all structure. Sinee the measurement of the parameters of an N-ter ‘inal-pair network can be reduced to the measurement of the parameters of two-terminal-pair networks, only the latter will be considered in this section. Sce.34] MEASUREMENT OF NETWORK PARAMETERS 131 ‘The three network parameters characteristic of an arbitrary two- terminal-pair waveguide structure can be measured by various methods. ‘The conventional network method involves the measurement of input impedancefor three particular values of output impedance. This method has tho advantage of being applicable to dissipative structures and to both ‘uniform and nonuniform lines. For nondissipative structures it is desir- able to employ pure reactive output impedances (short-circuited lines), since they give rise to an infinite standing-wave ratio in the input line. Under these conditions the output impedance is a simple function of the Iength of the short-circuited line, and the input impedance i a correspond- ly simple function of the distance to the minimum in theinput line. It is convenient to employ any three of the following pairs of measured values for input impedance Za and corresponding output impedance Zea, (at) for the determination of the unknown network parameters. ‘The quan- tity #o represents the input impedance set up by a zero output impedance, ete, In terms of the above values the impedance elements for a T repre- sentation (cf. Fig. 3-18a) of a two-terminal-pair network can be expressed. as in any of the following columns: ‘where the numbers at the head of each column indicate the particular set of three measured values in (81) on which the equations are based. ‘The admittance elements of a x representation (¢f. Fig. 8130) follow from the above expressions on the duality replacement of impedances by admit- tances. It is to be noted that either of two elements, + or —Zua(¥'), can be employed for the representation of the input-output impedance measurements. ‘This ambiguity can be resolved by # measurement of, transfer impedance or of any other quantity that yields the relative phase132 MICROWAVE NETWORKS (Seo. 84 at the input and output terminals. In many eases the correct sign inay be ascertained theoretically. For the ease of uniform lines the transformer representation discussed in See, 33 provides a basis for an alternative method of measurement of two-terminal-pair structure. As illustrated in Figs. 9-12 and 3-15 a nondissipative two-terminal-pair structure can be represented at charac teristic terminals "and 1? by an idesl transformer of impedance ratio n* connecting input and output lines af characteristic impedances Z, and Z, ‘The transformer ratio n* and the location of the characteristic terminals are readily determired from the following typical measurements of 1, The standing-wave ratio and position of the minimum in the input line for a matched load (Za. = Z) in the output line. 2. The position of a short cireuit in the output line such that the corresponding minimum in the input line coincides with that in ‘measurement 1 From the transformation properties of the ideal transformer it is evident that the standing-wave ratio in measurement. 1 is equal to n*Z/Z, which, as shown in Seo. 3:3, is denoted by ~7; the locations of the terminals 7} and T3 are given by the position of the maximum in measurement 1 and of the short circuit i measurement 2. Asin Fig, 3-12, the distances from the input and output terminals to TY and 7 are designated as Ds and Sq respectively. ‘Tae parameters of a 1’orx representation (cf. Fig.3-13a) are then expressed in terms of Ds, So, 7 by means of Eqs. (16) and (17). ‘The accuracy of the two preceding methods of measurement is diffi- cult to ascertain because of the uncertainty in individual standing-wave measurements. ‘This difficulty oan be partially removed by averaging a large number of such measurements. For the ease of nondissipative two-terminal-pair structures in uniform lines, a more systematic procedure may be employed if more accuracy is required. This precision method involves a plot of the measured values of the positions of the input minim: ‘vs, the correspondirg positions of the output short circuits. An anelys of this plot with the aid of the previously considered tangent relation in the form tan 2x(D! — Dj) = tan 2x(8" — 8) (32) yields the data required for the determination of the network parameters. Equation (32) is just Eq. (13) rewritten with D’ = D/Xx, Dy = Do/Aen 81 = S/don, and Sy = Sof. ‘As mentioned in Sec. 3-3, the tangent relation provides representa tion of the input-output impedance relation that is particularly well suited for measurements in nondissipative waveguides. ‘This is evident from Eqs. (14), whish indicate that S is identically the distance measured See. 84] MEASUREMENT OF NETWORK PARAMETERS 133 from the output terminals to a short circuit in the output fine and D is the distance from the input terminals to the corresponding minimum Position in the input line. The essence of the present method is the determination of a set of parameters Ds, So, and + that, on insertion into Eq, (82), provide a curve of D vs. $ which best reproduces the experi- mental curve of D vs. S. By Eqs. (15) and (17) it is apparent that the parameters Ds, So, and 7 are equivalent to the network parameters of a T or representation. A virtue of the present method is that a compari- son of the computed and measured curves of D vs. $ indicates immedi- ately the average acouracy of the final set of parameters Ds, Sy, and 7: ’ | o> 4 —s ! 1 1 So @ () ) In practice it has been possible to obtain, at wavelengths of about 3 em, ‘an average difference of less than 0.00054, between the experimental and, ‘computed curves of Dvs. 8. An accuracy of this magnitude implies that, the limitations in the accuracy of equivalent circuit measurements lie not in the standing-wave measurements but rather in the mechanical measurements required to locate the input and output terminals. ‘The details of a successive approximation procedure for the precise determination of the parameters Ds, Ss, andy will now be outlined. ‘The measured values of D’ and S' when plotted yield curves of the form indi- cated in Fig, 321.1 From Eq. (32) one sees that this curve should be repetitive with a period of a half wavelength in both D’ and S’ and sym- metrical about a line of slope —1. ‘The curve intersects the line of slope Lat points of maximum and minimum slope. ‘The point of intersection at the maximum slope is D’ = Dt, S' = 8%. The maximum slope is 7, and the minimum slope is 1/7. First approximations to Ds and S, are +A. Weistoch, Hochfre.u, Blektr., Vol. 60,1942, pp. 67 else.134 MICROWAVE NETWORKS (Sto. 34 ‘obtained from the ‘ocations of the points of maximum and minimum slopes by suitable averages. A first approximation to 7 is given by the average of the maximum and the reciprocals of the minimum slopes. An additional value for the average is given in terms of w/, the width in guide wavelengths Eetween points of slope —1, as won(-w) en If the guide wavelengths in the input and output lines are unequal, the value of w’ may be determined from a plot of D’ vs. 5’. In practice, however, it is most convenient to plot the absolute values of D and S as Fig. 8°21; in terms of the maximum spread w of this eurve ay (t_v fi, crewta(t-t frag) aw It is to be noted that the slope of the symmetry axis of this curve is ,/\a nd henee the maximum slope of this eurve is 7, Fig, 822,—Pt of D! +S! v8.8" for —y 261, Almost “matched” two-terminal-pair structures have a —y value of approximately unity and consequently give rise to a D vs. $ curve from which itis difficult to evaluate and locate the points of maximum slope. Since in such cases y = —1 — e(e 1), Eq. (82) may be rewritten in the approximate form D+ 8' 2D) + 8 — sin Ani’ — S). Bs) ‘Thus, if the experimental data is plotted in the form D’ + 5" vs. S', the carve shown in Fig. $22 is obtained. ‘The values of Dy + Si, Si, and 4/2 can be easily read from this curve and furnish first approximations to the required parameters Ds, So, and ‘The knowledge of the first approximations to Do, Ss, and y is sufi- She. 34] MEASUREMENT OF NETWORK PARAMETERS 135 ciently accurate in many cases. More accurate values can be obtai it theoretical curve of D vs. S, as computed by means of Eq. (82) and the first approximation values, is compared with the measured curve. A convenient mode of comparison is a plot of the difference ‘between the experimental and computed values of D vs, those values of 8 corresponding to the experimental points. This difference curve of AD = Dixy ~ Dioay V8. S' may or may not possess regularity. If regu- larity is exhibited, the first approximations to Di, Si, and y are inaccurate Drs tnn 02 ® ~0001 Fro, 323-—Plot of typialervor our (@) 7 = 2:85) = 0001, 3D, ay Ow) ee 'y = 001, ADs = a5, = 0 and the difference curve should be analyzed to obtain corrections ADS, ASs, and Ay to the first approximations. To determine these corrections itis necessary to know the expected form of the curve of AD’ vs. $' arising from variations ADS, AS, and Ay in Eq. (82). ‘The differential form of Eq. (2), namely, AD! = AD, ~ = m8 is the theoretical equation for the difference curve ‘The actual difference, or error, curve of AD! = Diy ~ Ding vs. 5° arises from errors ADs, AS, and Ay in the choice of Dj Ss, and y by the procedure described above. This curve is plotted in Fig. 3°23 for typical values of the vatious parameters. If, for example, the amplitudes of the actual error curve at $' — Sj = 0, §, and t are designated a8 As, Ay and 4, then from Eq. (35) the required corrections to be added to Ds, Sj, and ‘y are found to be136, MICROWAVE NETWORKS (See. 34 (7784 — Ao + (36) pg Ou = Ay Ay = Aal(L + 9) — do ~ 1*Oyd. Other methods of analysis of the error curve ean be employed depending on the value of 7. The corrected values for Dy, S%, and usually suffice to describe all the measured data to within experimental accuracy. ‘The accuracy itself can be estimated by plotting another difference curve ‘employing the second-approximation values of Dj, S%,and. This eurve should possess no regularity; its average deviation provides « measure of the average error in tho “eleotrical”” measurement. ‘The electrical ertor in the standing-wave measurements is to be dis- ‘tinguished from the “mechanies!” error in the measurement of distance tothe terminal planes, Since the evaluation of 7 and the location of the point of maximum slope ean be obtained merely from relative values of 1D and §, itis evident that these determinations involve only the elec- trical error. However, the absolute evaluation of Df and S{ necessary {or the determinatioa of the network parameters may involve the measure~ ‘ment of the distances from the point of maximum slope to the input and output terminals. Because of difficulties in maintaining accurate mechanical tolerances in microwave structures, the latter measurement js usually the larges: source of error. From the error curve, ete., it is possible to estimate the over-all experimental errors 4Ds, BS;, and By/y in Ds, 8%, and y. The corre- sponding relative errors 8a/a, 8¢/¢, and &b/bin the network parameters a, 5, fand¢ of. Bas. (18) arising from the experimental errors may be expressed ia [(_@a yy a = (eee, m8) + + (eRe) Fig *s) + (8): 8. (em) + GER) + (48 a) 7) ab 4rAy * 4wAs ¥ v Gets As i) + (GAS Ss; asi) + + (att 7)? where a@ tr) B+ ay!) Ager “=D tom) Ay = Bo + a) Ae =) * Tey — BR +a) = BG — a) Ape Vt) gat = 8) °” TE BG = a) (a= By) = ay) Ske. 34] MEASUREMENT OF NETWORK PARAMETERS 137 and a, 8, 7, a, b, and c are defined in Eqs. (15) and (17). ‘The sensitivity of the network parameters a, b, and e to errors in Ds, Se, and y evidently depends on the choice of terminals as well as on the type of microwave structure, For a desired accuracy in the values of the network param- eters, the above equations furnish from approximate values for Ds, So and 7 the precision necessary in the determination of the latter quantities. Instead of the relative error 36/b it is frequently necessary to know the relative error syora Vi Fac Distance Invariant Representations—As is evident from Eqs. (87) a precise determination of the equivalent-circuit, parameters of a micro- ‘wave structure requires a precise measurement of the distances Dy and Se, between the characteristic and the prescribed terminal planes. Accurate ‘measurements of these distances may require an absolute mechanical accuracy at \ = 3 cm, for example, of a mil in a distance of a few inches; this is exceedingly difficult to attain—particularly when the preseribed terminal planes are relatively inaccessible. Inaccuracies in mechanical measurements may result in disproportionately large erzors in the circuit parameters. For instance, inaccurate distance measurements on a sym- metrical structure may lead to an asymmetrical circuit representation (ie., ac) despite the certainty of structural symmetry. All such difficulties arise because of the dependence of the circuit representation fon “mechanical” measurements of distances to preseribed terminal planes. ‘The difficulties can be avoided by uso of a representation in which the type of cireuit is prescribed and in which the locations of the terminal planes are not prescribed but rather ascertained from the ‘measurements. ‘The values of tho circuit parameters in such a repre- sentation may be made highly acourate since they may be ascertained solely from the “electrical” measurement of the maximum slope + of the D vs. $ curve. Inaccuracies in absolute distance measurements have no effect on the determination of + and hence are manifest only 1s proportionate inaccuracies in the locations of terminal planes. ‘The determination of distance invariant representations of this type will now be discussed. ‘The equivalent circuit, representative of a nondissipative microwave structure is dependent on the determination of three parameters. As noted above, the tangent parameters De, S:, and form a convenient set. To secure a representation in which the cireuit parameters are inde- pendent of measurements of the two distance parameters Dy and So, it is necessary to prescribe two bits of information about the desired repre- (ta)138 MICROWAVE NETWORKS Bre, 35 sentation. For example, the circuit representation may be prescribed ‘to be both symmetrical and shunt. ‘The impedance parameters of the desired representation will therefore satisfy the preseribed conditions ae and (38) By Eqs. (15) this implies w=-l a=, (802) and hence a and are proscribed tobe 1 Va ‘The rlative eactance of the shunt element is therefore, by Ege. (36) and as), an tv—y 8 (390) omens 0) ao 0) ‘and is manifestly dependent on only the + of the D vs. $ curve of the given structure, The locations of the terminal planes for this shunt representation follow from measurements of the locations of the charac teristic reference planes and from the values [by Bas. (39b) and (17)] of Ds and Sq, the distances between the characteristic and the desired terminal planes. ‘There exist a variety of other distance invariant representations of microwave structure. For example, the representation may be pre- scribed to be purely series; the admittance parameters of this representa tion are then iden-ioal with the impedance parameters of Eqs. (40). ‘The transformer representation illustrated in Fig, 3-15 also belongs to this eategory. Of the various possible representations the most desirable is usually the one in which the associated terminal planes are located in closest proximity tc the physical terminal planes of the structure. 36, Theoretical Determination of Circuit Parameters.'—The pres- ence of a discontinuity structure in a waveguide results in discontinuities in the propagating mode fields at the “terminals” of the given structure. As noted in previous sections such field or, equivalently, voltage-current discontinuities can be schematically represented by means of a lumped- constant equivalentcireuit. ‘The theoretical determination of the equiva- lent-eizcuit parameters requires mathematical methods that do not properly lie within tie realm of microwave network engineering. Instead, such determinations generally involve the solution of so-called boundary value or field problems. ‘The present section is primarily intended to A comprehensive ecoount of the theory of guided waves is in proparstion by J. Schwinger and the author. Sec. 85] DETERMINATION OF CIRCUIT PARAMETERS 139 sketch those field theoretical techniques, devised largely by J. Schwinger, which have been employed to obtain the equivalent-circuit results pre. sented in Chaps. 4 to 8. ‘The field problems to be discussed are concerned with the behavior of electromagnetic fields not everywhere within a region but rather only in those regions relatively “far” from a discontinuity structure; the behavior in the latter is, of course, just that of the propagating modes. ‘The solution of such field problems presupposes the ability to determine tho electric and magnetic ficlds set up in a waveguide by electric currents, (ie., tangential magnetic fields) on obstacle-type discontinuities and by magnetic currents (ie., tangential electric fields) on aperture-type dis- continuities. ‘The field representations summarised in Eqs, (1-6) are of the desired form provided the mode funetions e,, the mode voltages Vs, and the mode currents I; can be determined. ‘The mode functions are so determined that the mode fields possess in the waveguide cross section the transverse ry behavior dictated both by the field equations (1-1) and. by the requirement of vanishing tangential electric field on the nondissipa- tive guide walls, Explicit evaluations of the mode functions e, for a variety of waveguide cross sections are presented in Chap. 2. The eorre- sponding evaluation of the mode voltages and mode currents then follows from the requirements that the mode fields possess the longitudinal 2 dependence dictated by the field equations, and in addition that the total fields satisfy the boundary conditions imposed by the presence of the discontinuity and the nature of the excitation in the waveguide. As shown in Sec, 1-2, the determination of the longitudinal z dependence of the mode amplitudes V; and [, constitutes a conventional transmission- line problem and is deseribed implicitly by the transmission-line equa- tions (1-8). ‘These transmission-line considerations are necessary preliminary to a major source of difficulty: the explicit evaluation of the electric or magnetic currents set up on discontinuity surfaces by the given excitation in the waveguide. These discontinuity currents must be so determined that the total fields satisfy prescribed boundary condi- tions on the discontinuity surfaces. Once the discontinuity currents are found, the various mode voltages and currents follow by straightforward transmission-line considerations. In particular there follow the domi- nant-mode voltage-current relations at the terminsls of the discontinuity, and hence the equivalent-circuit parameters characteristic thereof. ‘The preceding paragraph has sketched in only qualitative detail the salient features of a general method for the determination of equivalent- circuit parameters. The methods employed in Chaps. 4 to 8 are basically of a similar nature and differ mostly in their technique of successive approximation to the desired rigorous results. ‘These methods have been classified as140 ‘MICROWAVE NETWORKS [Seo 35 1. The variational method. 2. ‘The integral equation method. 8. The equivalent static method. 4. The transform method. ‘The particular methed employed in the derivation of the equivalent cir- cuit results in Chaps. 4 to 8 has always been indicated under Restrictions in each section therein. ‘The above methods will be briefly illustrated in this section. ‘The field problems presented by the Asymmetrical Capaci- tive Window desoribel in See. 5-1b (cf. Fig, 5-1-2) and the B-plane Bifur- cation described in Sec. 6-4 (cf. Fig. 6-4-1) have been chosen for simplicity to + _ : a Biected Equatont Longin Vinw ‘reat Yio, 524—Bieeted capacitive window. of illustration; the msthods to be disoussed are, however, of quite general applicability ‘The equivalent sireuit for a capacitive window in rectangular guide in which only the dominant Hyymode can be propagated is, in general, a four-terminal network. However, if the window is formed by an obstacle of zero thickness and if the input and output terminal planes are chosea coincident with the plane of the window, the equivalent net- ‘work becomes pure shunt, since the electric field and hence the dominant mode voltage are continuous at the terminal plane. ‘The associated field problem, whose solution is necessary for the determination of the relevant, equivalent-cireuit parameter, need be concerned with the field behavior in only the input half of the structure. This is a consequence of choosing the arbitrary excitation in the input and output guides such that the inward-flowing domizant-mode currents at the input. and output termi- nals are equal, Under these circumstances the tangential magnetic and clectrie fields in the aperture plane are zero and a maximum, respectively, with a consequent symmetry of the field structure about the terminal plane. A sketch of the bisected structure and its associated equivalent ircuit is shown in Tig. 3-24. It is first necessary to find in the input region 2 < 0 a soluticn of the field equations (1-1) subject to the boundary conditions of 1, Vanishing of the clectrie ficld tangential to the guide walls at ®. Suc. 85] DETERMINATION OF CIRCUIT PARAMETERS MI 2. Vanishing of the electric field tangential to the obstacle surface abs = 0,0

You might also like

• Equivalent-Circuit Models For

  

  100% (2)

  Equivalent-Circuit Models For

  5 pages
• Collin Antenna Theory Part 1 (BookFi)

  

  No ratings yet

  Collin Antenna Theory Part 1 (BookFi)

  682 pages
• High-Speed Signal Propagation Guide

  

  No ratings yet

  High-Speed Signal Propagation Guide

  12 pages
• Edward Conrad Jordan Electromagnetic Waves and Radiating Systems Prentice Hall Electrical Engineering Series 1968

  

  No ratings yet

  Edward Conrad Jordan Electromagnetic Waves and Radiating Systems Prentice Hall Electrical Engineering Series 1968

  774 pages
• Microwave Transistor Amplifiers (Analisis and Design) - Guillermo González

  

  100% (1)

  Microwave Transistor Amplifiers (Analisis and Design) - Guillermo González

  520 pages
• Collin

  

  No ratings yet

  Collin

  859 pages
• (0.5) A History of Microwave Filter Research, Design, and Development

  

  No ratings yet

  (0.5) A History of Microwave Filter Research, Design, and Development

  15 pages
• Slawomir Koziel, Xin-She Yang, Qi-Jun Zhang (Eds.) - Simulation-Driven Design Optimization and Modeling For Microwave Engineering-Imperial College Press (2013)

  

  No ratings yet

  Slawomir Koziel, Xin-She Yang, Qi-Jun Zhang (Eds.) - Simulation-Driven Design Optimization and Modeling For Microwave Engineering-Imperial College Press (2013)

  526 pages
• Solution Manual Antenna Theory by Balanis Edition2 Chapter14b

  

  No ratings yet

  Solution Manual Antenna Theory by Balanis Edition2 Chapter14b

  28 pages
• Transient Signals On Transmission Lines

  

  No ratings yet

  Transient Signals On Transmission Lines

  154 pages
• Wide Band Linearly Tapered Slot

  

  No ratings yet

  Wide Band Linearly Tapered Slot

  83 pages
• MIT Radiaton Lab Series V9 Microwave Transmission Circuits

  

  100% (2)

  MIT Radiaton Lab Series V9 Microwave Transmission Circuits

  737 pages
• Narrowband Microstrip Filter Design

  

  No ratings yet

  Narrowband Microstrip Filter Design

  48 pages
• Signal Processing Approaches To Secure Physical Layer Communications in Multi-Antenna Wireless Systems

  

  No ratings yet

  Signal Processing Approaches To Secure Physical Layer Communications in Multi-Antenna Wireless Systems

  142 pages
• Solution Manual Antenna Theory by Balanis Edition2 Chapter5B

  

  No ratings yet

  Solution Manual Antenna Theory by Balanis Edition2 Chapter5B

  16 pages
• Microstrip Antenna Design

  

  No ratings yet

  Microstrip Antenna Design

  434 pages
• Lecture 524 12 SIW

  

  No ratings yet

  Lecture 524 12 SIW

  20 pages
• Microwave Mixers PDF

  

  No ratings yet

  Microwave Mixers PDF

  143 pages
• Wireless Communication Electronics by Example: Robert Sobot

  

  100% (2)

  Wireless Communication Electronics by Example: Robert Sobot

  391 pages
• Time-Harmonic Electromagnetic Solutions

  

  100% (2)

  Time-Harmonic Electromagnetic Solutions

  25 pages
• Hybrid Coupler

  

  No ratings yet

  Hybrid Coupler

  14 pages
• Kraus Carver Electromagnetics

  

  50% (2)

  Kraus Carver Electromagnetics

  850 pages
• Antenna Radiation Pattern Guide

  

  100% (1)

  Antenna Radiation Pattern Guide

  5 pages
• Antenna Theory and Design

  

  100% (1)

  Antenna Theory and Design

  637 pages
• Ka-band LTCC Antenna Array for 5G

  

  No ratings yet

  Ka-band LTCC Antenna Array for 5G

  24 pages
• Time-harmonic Electromagnetic Fields (IEEE Press,) (491s) Harrington R.F. (高清版)

  

  No ratings yet

  Time-harmonic Electromagnetic Fields (IEEE Press,) (491s) Harrington R.F. (高清版)

  490 pages
• VSWR

  

  100% (1)

  VSWR

  27 pages
• Microwave Engineering 3rd Edition 2004 Pozar Solution Manual

  

  No ratings yet

  Microwave Engineering 3rd Edition 2004 Pozar Solution Manual

  288 pages
• 2017-Analysis and Design of Transmitarray Antennas

  

  No ratings yet

  2017-Analysis and Design of Transmitarray Antennas

  177 pages
• Ansoft HFSS Antenna Design Kit Design Parameters: Arien Sligar

  

  No ratings yet

  Ansoft HFSS Antenna Design Kit Design Parameters: Arien Sligar

  49 pages
• Metasurface Antennas For Synthetic Aperture Radar

  

  No ratings yet

  Metasurface Antennas For Synthetic Aperture Radar

  192 pages
• 2743 001 PDF

  

  No ratings yet

  2743 001 PDF

  30 pages
• Antenna Theory & Design Guide

  

  100% (1)

  Antenna Theory & Design Guide

  19 pages
• HFSS Scripting

  

  No ratings yet

  HFSS Scripting

  420 pages
• PPT

  

  No ratings yet

  PPT

  24 pages
• Oblique Incidence and Reflection Analysis

  

  No ratings yet

  Oblique Incidence and Reflection Analysis

  21 pages
• Applications of Modern RF Photonics

  

  No ratings yet

  Applications of Modern RF Photonics

  213 pages
• Signal and Power Integrity - Simplified - 2nd - Eric Bogatin - Prentice Hall PTR - 2010

  

  No ratings yet

  Signal and Power Integrity - Simplified - 2nd - Eric Bogatin - Prentice Hall PTR - 2010

  8 pages
• Ridge Waveguides and Passive Microwave Components PDF

  

  100% (1)

  Ridge Waveguides and Passive Microwave Components PDF

  342 pages
• RF MEMS Switches and Switch Circuits: Shimul Chandra Saha

  

  No ratings yet

  RF MEMS Switches and Switch Circuits: Shimul Chandra Saha

  174 pages
• On-Wafer Microwave Measurements and De-Embedding (Errikos Lourandakis)

  

  No ratings yet

  On-Wafer Microwave Measurements and De-Embedding (Errikos Lourandakis)

  251 pages
• B/C6D ,-&"/00 +, && $ce +fg./hi/,6j&6+ K&e/0 - ,& O:m&&p.c67 6ck&e/0 - ,&+Q&$G&7.C, 03 7&C, E&& ./6/ R/.&6+ K&C..CJ0&&

  

  No ratings yet

  B/C6D ,-&"/00 +, && $ce +fg./hi/,6j&6+ K&e/0 - ,& O:m&&p.c67 6ck&e/0 - ,&+Q&$G&7.C, 03 7&C, E&& ./6/ R/.&6+ K&C..CJ0&&

  6 pages
• B.N. Basu - Electromagnetic Theory and Applications in Beam-Wave Electronics-World Scientific Publishing Co. Pte. Ltd. (1996) PDF

  

  No ratings yet

  B.N. Basu - Electromagnetic Theory and Applications in Beam-Wave Electronics-World Scientific Publishing Co. Pte. Ltd. (1996) PDF

  492 pages
• Mathematical Foundations For Electromagnetic Theory

  

  No ratings yet

  Mathematical Foundations For Electromagnetic Theory

  2 pages
• RF Feedline - Theory and Problems of Transmission Lines (By PH.D Robert A. Chipman October 1968)

  

  No ratings yet

  RF Feedline - Theory and Problems of Transmission Lines (By PH.D Robert A. Chipman October 1968)

  244 pages
• Microwave Engineering Solutions Manual

  

  0% (1)

  Microwave Engineering Solutions Manual

  316 pages
• Radar Handbook - Scholnik

  

  No ratings yet

  Radar Handbook - Scholnik

  102 pages
• Microwaves and Radar Principles and Applications

  

  No ratings yet

  Microwaves and Radar Principles and Applications

  639 pages
• Microwave Engineering Akash

  

  No ratings yet

  Microwave Engineering Akash

  57 pages
• Carter Electromagnetic Waves - Microwave Components and Devices (Chapman & Hall)

  

  No ratings yet

  Carter Electromagnetic Waves - Microwave Components and Devices (Chapman & Hall)

  343 pages
• Rizzi Microwave Engineering Passive Circuits

  

  No ratings yet

  Rizzi Microwave Engineering Passive Circuits

  598 pages
• Electrodynamics for Engineers

  

  No ratings yet

  Electrodynamics for Engineers

  331 pages
• Unit 1 Part 1

  

  No ratings yet

  Unit 1 Part 1

  36 pages
• Electrodynamics & Transmission Lines

  

  No ratings yet

  Electrodynamics & Transmission Lines

  339 pages
• Ee 481 SDSMT

  

  No ratings yet

  Ee 481 SDSMT

  181 pages
• Syllabus: Prof ECE and Dean, R & D

  

  No ratings yet

  Syllabus: Prof ECE and Dean, R & D

  42 pages
• Microwave Engineering

  

  No ratings yet

  Microwave Engineering

  16 pages
• Microwave Engineering Lecture Notes

  

  No ratings yet

  Microwave Engineering Lecture Notes

  82 pages
• Microwave Engineering Fundamentals

  

  100% (2)

  Microwave Engineering Fundamentals

  649 pages
• Electromagnetic Waves and Transmission Lines - Machac

  

  100% (2)

  Electromagnetic Waves and Transmission Lines - Machac

  191 pages