US2656513A - Wave guide transducer - Google Patents

Description

A. P. KING n WAVE GUIDE TRANSDUCER oct. 2o, 1953 v2v Sheets-Sheet 1 'I Filed Dec. 29, 1949 mvg/fron A; P. KING'- 240 ATTOQNE Oct. 20, 1953 A. P. KING WAVE GUIDE TRANSDUCER 2 sheets-sheet 2 RFiled Dc. 29. 1949 BNL WY QF* Tl u v /a//Il /NVENTR Ap. ,fr/Na BV 7124 ATTORNE Patented Oct. 20, 1953 WAVE GUIDE TRANSDUCER Archie P. King, Red Bank, N. J., assignor to Bell Telephone Laboratories, Incorporated, New York, N. Y., a corporation of New York Application December 29, 1949, Serial No. 135,759

(Cl. S33- 21) Claims.

This invention relates to wave transmission` systems for the transmission of microwaves through hollow metallic tube Wave guides and more particularly to the conversion of guided waves from one mode to another.

Suchguided waves, as isv well known in the microwave transmission art, are capable of transmission in an iniinitely large number of forms or modes, each mode being distinguished by the characteristic configuration of the component electric and magnetic elds comprising the waves.

These waves have been divided into two broad classes. In one class the electric component of the wave is transverse to the metallic pipe guide,

, and at no point does it have a longitudinal component. The magnetic component, on the other hand, has both transverse and longitudinal components. This class has been designated as "transverse electric waves or TE waves. In the other class, the magnetic component is transverse to the pipe and at no point does it have a longitudinal component, but the electric component has in general both transverse and longitudinal components. This class has been designated as transverse magnetic waves or TM waves.

The waves in each of these classes have been Iiurther identihed and distinguished from each other by their mode or the pattern of wave energy distribution as it appears in the cross-section of the wave guide. A complete discussion of wave mode may be found in any standard textbook of microwaves and microwave guides. For the purpose of the present disclosure, the usual convention is herein adopted oi designating a transverse electric wave TEmn where, in a rectangular wave guide, m represents the number of half period variations of the transverse component encountered in passing across the width of the wave-guide cross-section, and n represents the number of half periods of transverse components encountered in passing across the height of the wave guide; and in a circular or cylindrical Wave guide, m represents the number of whole periods yof the transverse component encounteredin running around the circumference of the crossvof the guide. This is commonly known as the dominant mode wave.

Likewise, in a cylindrical wave guide, TEo1 represents a wave whose electric eld is wholly tangential and forms a series of circles concentric with the axis, and wherein no variations in the electric field are encountered around the circumference of the cross-section, but rather, a onehalf period is encountered along the radius thereof. known as a circular electric wave.

Each wave mode differs specically from other modes in certain transmission characteristics that are important with respect to the objects of this invention. For example, the circular electric wave, when propagated through a circular metal pipe of a given diameter, suffers progressively less attenuation as the frequency is increased. Because of its low attenuation the TEor wave is particularly adapted for long distance transmission systems. On the other hand, the IEro wave or dominant mode Wave is easily generated by presently known'generators, does not easily degenerate into other modes, and in addition, is most `favorable for amplication, modulation or demodulation.

It is therefore, an object of the present invention to convert wave power of a given mode -to wave power of another mode.

It is a particular object of the present invention to convert dominant mode wave power to circular electric Wave power and conversely circular electric wave power to dominant mode wave power by gradually varyingthe boundary conditions of the transmitted energy.

It is a further particular object of the present invention to convert 'IEm mode power to 'IEzo mode power and conversely 'I'Ezo mode power to 'IEio mode power.

It is a further object to permit the generation of a wave-mode which is especially favorable from a generation point of view and its conversion into a mode which is especially favorable from a propagation point of view by gradually varying the boundary conditions of the generated type wave.

Still a further object of the invention is to convert wave power of a given mode toksubstantially pure wave power of another mode by iiltering or absorbing therefrom spurious mode power having an undesirable form..

In accordance with the objects of the invention, wave power of a given mode is convertedin a wave transducer section to wave Y power of another mode by gradually varying the boundary conditionsof the transmitted energy by varying the geometry of the transducer section.

In a iirst embodiment of theinvention, to be This latter wave mode is commonlyv hereinafter described in detail, dominant wave power TEM) in a rectangular wave guide is converted into circular electricTEol wave power by bifurcating, or dividing into two portions, the TEN wave in the rectangular wave guide, launching it into two wave guides each having approximately half the height of the original, rotating the two portions counter-clockwise and clockwise, respectively, sectorially expanding each of the rotated power portions in a gradual manner to obtain a total circular configuration, and then recombining the separate portions in a circular wave guide.

A particular feature of the invention resides in an alternate biiurcating and rotating portion of the transducer, wherein the TE@ wave power is applied to the mid-point of 'a ISO-degree waveguide bend by means of a T connection in the plane of the electrical field. The energy is bifurcated and brought out in two side-by-side wave-guide sections, each equal in time phase but opposite in space phase.

In a second embodiment of the invention, dominant wave power T310 in a rectangular wave guide is converted directly into circular electric wave power TEoi in a circular wave guide by gradually varying the boundary conditions of the transducer section from rectangular cross-section tol a circular sector cross-section, and then expanding the sectoral cross-sectionwaveguide with gradually increasing sector angle until said angle reaches 360 degrees and the transition to circular electric wave is completed.

Another feature of the invention resides in the fact that the herein disclosed transducers are bilateral devices, i. e., energy may be transmitted aand converted in either direction. While, for the sake of simplicity in the following detailed description, the embodiments will be shown and described as having input and output terminals, it should be borne in mind that the converse transmission applies in each case also.

A further feature of' the invention resides in the spurious mode wave absorber which substantially dissipates the power of the spurious mode waves produced in the transducer section.

` The nature of the present invention, its various objects, features, and advantages Vwill appear more fully upon consideration of the embodiments illustrated in the accompanying drawings and the following detailed description thereof.

In the drawings: l U

Fig. 1 represents 'a TEM to TEM wave trans'- ducer; i* y Figs. 1A through IH, J, andIK represent cross-sectional views of Fig. i as indicated;

Fig. 2 shows an alternate embodiment of the transducer of Fig. l; v Y

Fig. 3 shows in detail a component of the structure of Fig. 2; V Y A Fig. 4 illustrates a single cavity 'IEit to TEm transducer; y

Figs. 4D through 4H, 4J and 4K represent cross-sectional views oi Fig. 4 as indicated;

Fig. 5 shows in detail the spurious Inode wave absorber incorporated in Figs. 1, 2 and 4;

Fig. 5A shows a cross-section view of 5 as indicated; and l l A Fig. Ao shows in detail a portion of Fig. 1 or Fig. 4.

Referring now to the transducer shown in Fig. l, a dominant TEN; wave to be converted into circular electric wave power is applied to one end of rectangular wave guide II. At the other end the wave power is bifurcated into wave guides I2 and I3, and then applied to twisting members I4 and I5. Members I4 and I5 which are connected to guides I2 and I3, respectively, are progressively rotated between section B and section D in a manner to be described. The wave energy from guides I4 and I5 is applied at section D to sectorial section It, twist member I4 delivering its energy to cavity Il, and member I5 delivering its energy to cavity I8. Cavities rIl and i8 expand the wave front with increasing sector angle between sections D and H until the combined angle of the two cavity sectors reaches 3,60 degrees at section H. At section H the cavities are separated only by a thin partition i9. Energy appearing at section H is applied to the spurious mode wave absorber 50, the function and construction of which will later become apparent in connection with Fig. 5.

Wave guide ii is of the conventional and familiar rectangular hollow pipe guide construction having a height b measured along the electric vectors and a Width de measured across the electric neld. It may be made of any electrically cohducting material and, as is usual in the art, may be nlled rwith gas or any other dielectric material, or if preferred, evacuated. Bifurcating guides I2 and I 3 are connected to wave guide Ii so that microwave energy present in the input guide I I is divided into two equal portions in guides I2 and I3, respectively. Guides i2 and I3 are each of approximately half the height and substantially equal the width of wave guide II so thattheir combined cross-sectional area is equal tothe cross-sectional area of guide il. This provides a good impedance match between guides I2 and It and the input guide I I.

The twist section, made up of members I4 and I5, extends from section B to section D. Members I and i5 are rectangularly cross-sectioned metallic waveguides having interior cross-sectional diniensions equal to guides I2 and I3, respectively. Each of these rectangular guides is provided with a linear twist or rotation about the longitudinal axis of the guide. The total of this twist is degrees, distributed linearly along thelength of the guide from section B to section D. i Thus the cross-section of each guide at one end in section B is displaced 90 degrees from the cross-section at the other end in section D, and at a point mid-*way between sections B and D, the cross-section of each of the guides will have been rotated 45 degrees as shown in Fig. 1C.

-These twisted members It and I5 are connected at one end of each to guides I2 and I3, respectively, at section B, so that member I 4 twists in a clockwise direction and member IB in a counter-clockwise direction when viewed from section B; In order to provide proper separation between the twisted member at section D, guides I2 and i3 are provided with a gradual S curve from the input end to section B. At section B the cross-sectional openings of guides I2 and I3 appear in the saine plane. l Thus the two portions of microwave energy impressed in phase upon guides I4 and I5, as indicated by the electric vectors in the same direction on Fig. 1B, will each be rotated along the length of the twist section until the two portions of energy are equal in time phase, but opposite in space phase, at section D as indicated the electric vectors which appear in opposite directions on Fig. 1D; If, therefore, a wave guide having a height equal to that of guide I4 and a width approximately twice that of guide I4 were integrally connected to the twist members acume at section D, a TEzo wave would be launched therein. Rather, these two portions of oppositely polarized wave energy are impressed on sectoral cavity section I6, guide I4 delivering its energy to cavity Il and guide I5 delivering its energy to cavity I8. At section D each of cavities I'I and i3 has a rectangular cross-section of the same size and shape as guides I4 and I5 to which they are connected.

Between section D and section E the transverse cross-section of each of the cavities is gradually changed from rectangular t0 sectorial. One characteristic which is ideally desirable for this section is that the cross-sectional change thereof be made at a constant impedance, and for this purpose, complicated mathematical expressions may be derived which would specify the exact dimensions of the cross-section at successive points between section D and section E. However, if the distance between the sections is at least several wavelengths, it is satisfactory to design the cross-sections at each end of the section to have equal impedances and to vary the dimensions of the cavity therebetween in a gradual and linear manner.

An enlarged section of either one of the cavities I'I or I8 between the rectangular cross-section at section D and the circular sector crosssection at section E is shown in Fig. 6. As indicated above, one characteristic of this section, the design for which will now be considered, is that the characteristic impedance of the rectangular section at D should equal the characteristicimpedance of the circular sector sectionat E. The

characteristic impedance of the rectangular section at D is given by the following expression:

the sectorial section at E is given by the following expression:

where 0 is the angle of the sectorin radians and a is the radial dimension thereof. In many cases some of the factors in the above equations may 'be determined by practical considerations. For example, the dimensions ar and b may be fixed by the standard size of the rectangular wave guides of the system in which the transducer is to be used. Likewise, a may be determined by the radius of the circular wave guides of the system, out it should be noted that a and 0 may be varied arbitrarily within certain limits to obtain a given value of sectorial section impedance. It is therefore possible to equate Equations 1 and 2 and to solve for the proper value of sector angle which would make the impedances at each end of the section equal. .In substantially lall waveguide systems, 9 will be an acute angle. With the proper value of the sector angle chosen, the dimensions of the cavity are varied graduallybetween the rectangular section YD and the sectorial section E. The outside width dimensions of each cavity, indicated on Fig. 6 as is linearly increased in length and at the same time arched or curved to form the arcuate wall of the` sectorial cavity. At the 'same time the inner dimensions, indicated on Fig.v 6 as y, is decreased at a uniform rate between sections D and E, becoming zero at E, to for-m in effect the vertex of the sectorial cavity. The transition in cross-section shape is made graduallyvin order that small impedance variations which might occur between section D and section E are suiciently distributed so that the eiect of mismatch appearing at the terminals is low.

n It is also possible to make the phase velocity of a wave at the rectangular section equal to that at the sectorial section, while at the same time keeping the impedances of the sections equal. The phase velocity atthe rectangular section is given by the expression:

VD D* 1v (a.)

where C is the velocity of light in centimeters.

The phase velocityat the sectorial section is given by the expression:

Equations 3 and 4 may be equated and the proper dimension ratio of will be found to be equal to V1.22. With this ratio established between a and ar', Equations 1 and 2 are employed in the manner already described to obtain the proper value of 0 which ,will make the characteristic impedances equal. Both conditions, i. e.,requal impedance and equal phase velocity, are important for maximum eflicienoy of the transducer section, but if, in a particular embodiment, the designer is limited by standardized Wave-guide component dimensions, it is substantially moreimportant that the i-mpedance qualifications 'be met than thatthe phase velocity at the two sectionsbe made equal. i

With attention directed once more to Fig. 1, it is noted that the cross-sectional change described with reference to Fig. 6 takes place on both cavities I'I and I8 in substantially the same manner between sections D and E. Thus, at section E the cavities appear in the cross-sectional view of Eig. 1E as two sectorial sections each having a sector angle 0 and separatedfrom each other bythe vertex partition' 2 I.

Between sections E and `H the sector of each cavity is progressively expanded in a linear manner along the length of the section. Thus, the angle 0 of each cavity. is increased from the acute angle at section E, in proportion to the distance `cross-sectional view taken at the point. The

, phase front of the wave is expanded'jin eachsection by the sectorially 'expanding boundaryconmetallic block 30,

dition, f the cavity; The component-.sof electric field adjacent the arcuate wall of each cavity are increased in length las the sector angle is increased. At section H the transition is complete 'and the field distribution of a TEoi circular electric wave has been duplicated in the two portions of the wave power, each expanded in cavities Ii and I3, respectively. These two portions are separated now only by the radial partition I9 as seen in Fig. lI-I. In the section immediately following section H, this partition no longer exists, and the two portions combine to form a circular electric wave. Since the iield distribution of the sectorial waves in sections E, F and G is the same as that for the TEU; wave in section H, the transition to TEni wave power is electrically smooth.

In spite of this fact, it is evident that the individual rays of the microwave energy will follow spiral paths in expanding the sectorial wave to the completed TEoi wave, and that the path length from a point in section E corresponding to a wave front point in section H is not equal for all path lengths. Such a difference in path lengths will inherently introduce other types of wave forms, for example, TEn and TEzi, in addition to the desired TEoi wave. This introduction of spurious modes is minimized by making the transitions from one cross-section to the next gradually and linearly. It remains, however, to filter out the remaining spurious mode power from the microwave output of the transducer if pure TEoi wave power is to be obtained. For this purpose the spurious mode wave absorber t is connected at section H to the cavity transducer.

A complete explanation of the spurious mode wave absorber will be reserved for later treatment in connection with Fig. 5.

Fig. 2 shows an alternate arrangement of the transducer of Fig. l wherein element 30, as shown in detail in Fig. 3, replaces the bifurcating and twisting members II, I2, I3, I4 and I5. Element 30, therefore, must perform the function f translating the total TEio wave into two portions of oppositely polarized wave power which, if combined, would form a TEzo wave. The remaining elements, the cavity section I6, which transforms the oppositely polarized'portions into TEoi wave power, and the spurious mode wave absorber 53, are identical to the corresponding elements shown in Fig. 1.

The operation of transducer 30 will become apparent upon consideration of the detailed section shown in Fig. 3. In Yits basic conguratiori the tranducer comprises a A18o-degree wave-guide bend with aninput llil-plane T connection at the center point of the bend. A section o f rectangular wave guide having opposite end openings 35 and 36 is bent 180 degrees so that the ends 35 and 36 appear side by side in the same transverse plane. A rectangular input guide 3I opens into the top face of guide 35-35 at the mid-point of the IBG-degree bend by means of an lil-plane T connection. The crossesectional dimensions of guide 3| are equal to those of guide II of Fig. l, which it replaces, and the cross-sectional dimensions of guides 35 and 3S should be equal to those of members I4 and I5, respectively, which they replace. This means that the width, as measured across the electric eld, of guides 3l, 35, and 36 are equal; while the height of guides 35 and 36 is one-half that of guide 3 l As shown in Fig. 3, the transducer is integrally constructed with the. basic wave-guide components described above machined or cast in a The bend portion oi guide 7.35e3 vbecomes a semicylindrical cavity having a semicylindrical wall 3 3. The straight guide sections 35 and 36 extend side by side perpendicularly from the diametral face of the cavity, and

.each have a height equal the height of the cavity and a width approximately equal the radius of Yany arbitrary shape, so chosen to provide the proper impedance characteristics.

Input guide 3l, having a connection flange 32,

is connected to the upper face of the cavity with its i:heligliit dimensnn perpendicular to the diame ra ace o t e cavity and approximate] equal to the radius thereof. y

The incoming TEic wave to be converted into circular electric wave power is applied to rectangular guide section 3| with an electrical field direction as indicated by the arrows on Fig. 3.

The wave propagates down the guide to the point of juncture with the semicircular cavity. At this point the conductor ends, and the electrical eld begins to fari out or expand in an attempt to conform to the new boundary conditions of the cavity. This expansion of the electric field is indicated by the arrows on Fig. 3. In this state of field distortion, the power is bifurcated by parti- `tion 39, one-half the Vpower going into guide 35 and the other into guide 36. The portion in each half is of opposite electrical polarity. As shown on Fig. 3, the vector indicating the electric field direction in guide 35 points downward while the vector in guide 36 points upward. These two portions of oppositely polarized TEio wave power are brought out side by side, and, as was the case of the power delivered by twisting members I4 and I5, would combine in a single wave guide connected by means of ange 31 to the openings of 35 and 3B to form a TEzo wave. Rather guides 35 and 36 are connected to cavity section I6 so that the guides deliver their power into the cavi- I ties I'I and I8, respectively.

Thus the transducer of Fi 3 same function as the structuregof Fipgroarelwg section A and section B, comprising elements I I I2, I3, I4 and I5. Transducer 3B is, however ci' much more rigid construction and may be built with higher precision. It is thus possible to achieve a higher amplitude and phase equality between output wave guides 35 and 3G than betwen tjlrile twist members I4 and I5 and thus re uce e amount of at terminals. dominant wave appearing The type of wave transducers descr' fore with reference to Fig. l throughlgiglrlelgeifes satisfactory performance in the centimeter wavelength band, At these wavelengths it is possible to match the two branches of the transducers producing oppositely polarized portions of wave power moderately well with respect to amplitud and phase characteristics. It is more diiiicult3 however'z to realize the higher precision required 0 obltain idientical construction of the two ranc circui s for fre u range. q encies in the millimeter The single-cavity wave transduc Fie. 4 avoids this difncuity of matciirngsrtintcifl branch circuits by exciting a single cavity directly with TEio wave power at input 4I. The trans'- 9 Vducer consists essentially of two parts: the rst part, cavity 42 extending from section D to E for converting the TEm wave power to sectorial TEM wave power in which the boundary conditions are varied gradually from the rectangular cross-section of D to the circular sector cross-section of E; and the second part, comprising cavity 43 extending from section E to J in which the boundary conditions of the sectoral wave cavity 43 are expanded until the sector angle 6 reaches 360 degrees. the circular electric wave energy is applied, as in Figs. 1 and 2, to spurious mode Wave absorber 50.

The first part comprising cavity 42, extending from section D to section E, may be identical in its essential design characteristic to the cavity shown in Fig. 6 and hereinbefore described in detail. The input opening 4I has a transverse rectangular cross-section in section D having a height b and a width ar and has a characteristic impedance as given by Equation l. Between section D and section E the cross-section of cavity 42 is gradually changed in the manner described in relation to Fig. 6, until at section E, the crosssection becomes a circular sector having a radius a and a sector angle 0. The circular sector opening thus has a characteristic impedance given by Equation 2. The value of the sector angle is so chosen in accordance with Equations 1 and 2 to make the characteristic impedance at section E equal to that at section D. Further, it may be desirable to make the phase velocity of the Wave At section J passing section D and section E equal by choosing a sector radius equal to 1.22 times the width dimension of the rectangular cross-section at section D as shown by Equations 3 and 4. The longitudinal distance between sections D and E should be at least several wavelengths and the variation in cross-section is made linearly along this length.

Between section E and section J the sector angle 0 is linearly increased from an acute angle to a plane angle. Thus, the acute angle sector at section E becomes a plane angle sector or a circular sector having a sector angle of 360 degrees at section J. At each successive cross-section between sections E and J the sector angle 0 increases in relation to the distance of the section along the longitudinal axis of the transducer. Several successive cross-sections -are shown in Figs. 4F, 4G and 4H. As the sector angle is increased, the boundary conditions of the cavity are changed, and the wave front of microwave energy passing through the cavity is sectorially expanded. In terms of the lines of the electric field of the wave this means that the lengths of the lines adjacent to the arcuate wall of the sectorial cavity are increased as the wave progresses down the length of the transducer more than the lengths of the lines near the vertex of the cavity. Due to they attempt of the electric eld lines to conform to the boundary conditions of the cavity, the lines become concentric about the cavity vertex. This is the field distribution of the circular electric or TEM WaVe.

The relative purity of the generated TEM wave is a function of the geometry of the particular transducer design and its dimensions in terms of wavelength. In considering the latter, the diameter of the TEM section determines the number of TE and TM waves which can be supported by the guide at the operating frequency. The total number of modes possible in a circular l wave guide varies as the square of aA, where ar 1s the radius of the wave guide and A is the operating wavelength. For example, a wave guide with a, radius of 1 inch and conducting energy with a wavelength of 6 millimeters, a value tentatively considered as a practical region of operation in a TEM Wave communication system, can support about spurious modes.

The number of spurious modes possible is related to the TEM wave transducer design problem in the following way. In transducers of the types described above, the mechanism for expanding the wave front involves diiferent path lengths from various parts of the sectorial TEM wave front to a transverse plane at the TEM end of the transducer. This feature has been pointed out in connection with Fig. 1. The result is that the equiphase surface at the TEM end is not a transverse plane but a curved surface. Such an equiphase may be represented as a mixture of waves, TEM-I-TEmn-l-TMmn Waves. In general a given amount of wavefront distortion or curvature will generate a greater number of spurious mode waves as the guide is operated further from cut-off, with an increasing ratio of a/A.

On the basis of the path lengths of the individual rays which follow spiral paths in expanding to the sectorial TEM wavefront, the variation in length or phase may be calculated. In the case of the single-cavity wave transducer of Fig. 4, it may be shown that the variation in the path length relative to the plane of section J depends inversely upon the length L of the cavity from section E. to section J, or that portion of the transducer which expands the wavefront with an increasing angle 0. If L is made several wavelengths long, or many times the radius a of the sector, the magnitude of the spurious modes generated may be maintained small compared with the magnitude of the desired TEM wave. This applies equally well to the corresponding Wavei'ront expanding portions of the transducers of Figs. 1 and 2. y

It may be desirable to further decrease the spurious mode power in the TEM circular wave output. For this purpose a further section, which has been referred to as the spurious mode wave absorber, is incorporated into the wave transducers. This section is indicated on Figs. 1, 2 and 4 as spurious mode absorber section 50, which eliminates, the spurious modes by absorbing them.

Referring to Fig. 5, the spurious mode wave absorber is shown in detail. It comprises a plurality of thin strips of resistive material, as 5| and 54, placed radially in the circular waveguide section 53. These strips may be self-supporting or they may be supported by means of polyfoam or styrofoam spacers, as 52 and 55, in the wave guide.

While the spurious mode wave absorber is of primary importance in absorbing spurious modes generated by a circular electric wave transducer as shown and described herein, it is by no means limited to this application. spurious modes may arise as the result of an imperfect transmission line due to small deviations from circularity or slight impedance irregularities in the line. Further, they may be developed by terminal devices, tapers, or other components which introduce either by reflection or by transmission, modes other than the desired one. In each of these cases the spurious mode powerrmay be suppressed by the use of the absorber employed in conjunction with the device tending to produce the undesired modes.

section is rst realized being substantially wherein b is the smallest dimension of said rectangular cross-section and A is the wavelength of said electric wave power, whereby said acute angle section has a characteristic impedance substantially equal tosaid given characteristic impedance, and the successive cross-sections of said enclosure from said point to said second end varying smoothly as the length of said arcuate wall increases in relation to the distance of the successive cross-sections from said point.

3. The combination according to claim 2 wherein the sector angle of said successive crosssections of said enclosure from said point to said second end varies gradually from said acute angle sector to a straight angle sector at said second end.

4. The combination according to claim 2 wherein the sector angle of said successive crosssections of said enclosure from said point to said second end varies gradually from said acute angle sector to a plane angle sector at said second end.

5. A wave transducer for converting incident electric wave power from one mode into another mode comprising an enclosure of electrically con ducting material forming a semicylindrical cavity having top and bottom walls, a diametrical plane, and an arcuate wall, a Wave guide opening into said cavity on said top wall for receiving said incident wave, and -a pair of contiguous rectangular wave guides opening into said cavity on said diametrical plane, a pair of elongated electrically conducting enclosures each having a rst end radians,

14 and a second end, said first end of each of said enclosures being connected respectively to one of said pair of rectangular wave guides, the transverse cross-section of each of said enclosures being rectangular at said rst end and having a given characteristic impedance, said cross-section of each of said enclosures varying from said first end cross-section to an acute angle sector cross-section at a point between said first and second ends, said last-named cross-section having a characteristic impedance equal to said given characteristic impedance, the successive crosssections of each of said enclosures from said point to said second end varying smoothly along the length of each of said enclosures with said sector angle becoming successively greater than said acute angle in relation to the distance of the successive cross-section from said point, and means for combining said two portions in a circular cross-section,

ARCHIE P. KING.

References Cited in the file of this patent UNITED STATES PATENTS Data, by Moreno. Published by Sperry Gyroscope Co., Great Neck, Long Island, N. Y. Received in Patent Office February 18, 1946, pp. 69 and '70. Copy in Div. 69.

Images