US2759099A - Plural-source coupling arrangements - Google Patents

Description

Aug. 14, 1956 5. A. OLIVE 2,759,099

PLURALSOURCE COUPLING ARRANGEMENTS Filed May 20, 1954 2 Sheets-Sheet 1 71 0 Past/2,9770! IN VEN TOR.

United States Patent PLURAL-SOURCE COUPLING ARRANGEMENTS George A. Olive, Princeton, N. .L, assignor to Radio Corporation of America, a corporation of Delaware Application May 20, 1954, Serial No. 431,167

8 Claims. (Cl. 250-36) This invention relates to arrangements for coupling a plurality of high-power sources of radio frequency energy to a common load, without interaction between the sources. This coupling is effected without causing a damaging flow of power from the high-power sources to a low-power radio frequency source which injects power into the high-power sources, for frequency stabilization purposes.

An object of this invention is to provide a novel type of wave propagation device, particularly useful in the microwave region of the frequency spectrum and capable of handling high power.

Another object is to provide a novel diplexer or power combining device by means of which the outputs of two transmitters may be combined in a common load, without interaction between the transmitters.

A further object is to devise a novel arrangement for the injection locking of magnetrons, by means of which the source of injection power is protected from damage by the magnetron power.

A still further object is to provide an arrangement for operating two injection locked magnetrons in parallel, in a manner which protects the source of injection power.

The objects of this invention are accomplished, briefly, in the following manner: A length of circular waveguide is provided with means of coupling to the dominant mode of wave propagation therein, at four points. These coupling means may for example consist of coaxial lines having probe-ends projecting into the waveguide. Two of these coupling means are near one end of the waveguide and are orthogonally positioned with respect to each other, and the other two coupling means are near the opposite end of the guide, are orthogonally positioned with respect to each other, and are angularly displaced with respect to the first two. When this device is used as a diplexer, two sources of radio frequency energy (transmitters) are coupled to the respective two coupling means near one end of the guide and powerabsorbing devices are coupled to the respective two coupling means near the other end of the guide. When the aforesaid device is used for operating two injection locked magnetrons in parallel, the two magnetrons are coupled to the respective two coupling means near one end of the guide, by way of transmission lines of different lengths, while a load and an injection power source are coupled to the respective two coupling means near the other end of the waveguide.

The foregoing and other objects of this invention will be better understood from the following description of some exemplifications thereof, reference being had to the accompanying drawings, wherein:

Fig. 1 is a schematic illustration of a single-section circular waveguide diplexer;

Figs. 2, 3 and 4 are combined vectorial and structural diagrams illustrating certain features of operation of Fig. l;

circular waveguide diplexer used for power combining; and

- Fig. 6 is a schematic illustration of a double-section circular Waveguide diplexer used for the injection locking of two magnetron oscillators.

The demand for increased transmitted power for television service, along with the practical necessity for feeding power from both the video and audio television transmitters into a single antenna, has brought about a need for high-power apparatus for diplexing the outputs of two transmitters into a common load. A diplexer for this use should also operate to decouple the two transmitters from each other, since it is obviously undesirable to have crossfeed from the video into the audio transmitter, or vice versa. The diplexer of this invention satisfies both of the aforesaid requirements.

The demand for increased power at ultra high frequencies for television service can be met by the use of high-power continuous wave (CW) magnetrons, such as those which will give an output of 15 kilowatts at 50% efiiciency. For use as a television transmitter, the anode voltage of the magnetron is modulated with superimposed alternating voltage representing the intelligence. However, this variation in the magnetron anode voltage results in a varying output frequency which must be corrected before the device can be used to provide an ampli- Fig. 5 is a schematic illustration of a double-section tude modulated carrier. At lower power levels, such as the one-kilowatt level, the magnetron frequency can be satisfactorily locked to a crystal-controlled signal injected at the magnetron output loop, as disclosed and claimed in the copending Koros application, Serial No. 177,455, filed August 3, 1950, and since abandoned. The performance of such an injection locking system regarding the depth of modulation and the amount of incidental phase modulation, has been found to be satisfactory for television transmission. However, when higher-power magnetrons, and particularly two high-power magnetrons operating concurrently and in parallel, are injection-locked by the above-described arrangement to operate synchronously and in parallel, ditficulty is experienced in obtaining adequate isolation of the injection power source from the high magnetron power. A different approach to effect the locking consists in the use of a combining network which provides locking signal to the magnetrons, isolation of the magnetron power from the injection power source, and addition of the two magnetron outputs in the common load. The requirements of locking, isolation and power combination are satisfied by the combining network of this invention, which is a type of microwave diplexer. Also, the diplexer of this invention is capable of handling very high power.

Now referring to Fig. 1, a single-section circular waveguide diplexer is shown, including a length 1 of circular waveguide wherein the dominant wave mode is propagating and two side arms 2 and 3 the axes of which are at right angles to each other as indicated, that is, these arms are orthogonally positioned with respect to each other. For ultra high frequency work, the side arms 2 and 3 may be coaxial lines with the respective inner conductors 4 and 5 penetrating into the circular guide 1 and serving as probe-ends for launching waves into or extracting energy from the guide. For reasons that will become apparent hereinaften the side arms 2 and 3 are axially spaced, and in particular the axial spacing between the two orthogonal arms 2 and 3 should be an exact multiple of a half-wavelength at the operating frequency in the guide. For purposes of simplification, this spacing is taken as one guide wavelength k (which is of course a multiple of a half guide wavelength), so that the two side arms may be considered to be located in the same transverse plane.

The inner diameter'of the circular waveguide 1 is below the cutoff dimensional value for all wave modes except the dominant mode at the operating frequency. For Hn-mode in the circular guide, the cutoff wavelength 7m isexpressed by the relation where a is the radius of the waveguideand 1.84 is the where x is the free-space wavelength.

The mode transmitted in the coaxial arms Z-and 3 consists of principal (TEM) waves and this mode is transformed into H11mode 'at'the junctions between the coaxial'arms and the-circular waveguide. The transition region between each coaxial-arm and the guide acts as a mode transformer. For proper rnatching, there exists an optimum'ratio of the diameter of the probes to that of the circular "guide, and an optimum depth of penetration of the coaxial inner conductors (probes) '4 'and into the guide 1, foranypa'rticular impedance of coaxial line.

The properties of coupling and of isolation in the side arms 2 and 3 of the circular waveguide diplexer illustrated in Fig. 1 depend on theorientation of the polarization plane of the field in thecircular guide. More par ticularly, in order for a'wave in the main guide 1 to be coupled to either side arm 2 or 3, theelectric vector representing the wave must have a component parallel to the longitudinal axis of that side arm. The degree of coupling between them depends on the magnitude of this component. This will now be further explained with reference to Figs. 2 and 3.

It will be recalled that'the two side arms 2 and 3 may be considered to be located in the same transverse plane, since they have an axial spacing of one guide wavelength. In Fig. 2 the electric field vector in the circular guide '1 is perpendicular to arm 3 but is parallel to arm 2,.resulting in zero coupling-with arm 3 and maximum coupling with arm 2. In Fig. 3, the electric field vector is at equal angles with both arms '2 and 3. Equal coupling is thus provided to each of these arms.

Since the plane of polarizationofthe :waves excited by arm 2 is perpendicular to that excited by arm 3,'it may be seen that power'incident in one side arm cannot theoretically couple into the other arm. In practice, to avoid direct coupling between theside arms, such arms should. be axially separated asufiicient distance (for example, a wavelength in the circular guide) and the structure should be carefully matched from an impedance standpoint.

In Fig. 1, arms 1' and 1"form the ends of the length 1 of circular waveguide. If the plane of polarization of the electric field vector in arm 1' at one end of the circular guide is at right angles to that of arm 1", as shown in Fig. 4, these two arms are isolated from each other. They both couple to arms 2 and '3, however, for values of 0 (see Fig. -4) between zero and ninety degrees. .A wave incident on arm 2, setting up E2 in the circular guide, couples Eacos a'into arml and +112 sino into arm 1"; similarly, and incident wave in arm 3 (E3 in the circular guide) couples E3 sin 6 into-arm 1'- and .E3 cos 0 into arm .1". Arm 1' is the combination arm and arm 1" is the difference arm. By a proper choice ofthe angle 0 and the magnitudes of the fields, the wavesfed to arm 1" may be made to annul and those fed to arm 1' to add. Reversing the direction of either E2 or'Ea will changes the role of arms 1' and 1". Conversely, if arms 1' and 1" are made the input arms, similarcombination and cancellation properties can be found in the output arms '2 and 3.

In connection with ultra high frequency Work, it is not convenient to use a circular guide for the input or output arms which are represented by 1' and 1" in the single-section diplexer of Fig. l. The system may be modified for external connection to coaxial lines throughout by joining two single-section diplexers together, as illustrated in Fig. 5. In actual practice, the structure may consist of two diplexers of the type of Fig. l placed end-to-end, one being rotatable with respect'to the other (the construction illustrated in Fig. 5), or of a single length of circular waveguide having'four side arms with proper angular and axial spacings. The former construction allows adjustment of "the angle 0 between the input and output sections, and also provides some degree of flexibility in correcting 0 to the proper value when necessary mechanical tolerances have not been maintained in manufacturing.

In Fig. 5, the input circular waveguide section 1 and the output circular waveguide section 6 are placed endto-end, with'a rota-table coupling therebetween. In waveguide section. 1, the two coaxial lines 2 and 3 havetheir axes perpendicular to the'axis of the guideand are arranged to'actas coupling means to the guide, just as in Fig. 1. These lines 2 and 3 are at with respect to each other, .as indicated, thus being orthogonally positioned with respect to each other. The free end of input section 1 is closed by a piston, as is also the free end of output section 6. In output waveguide section 6, which like input section 1 is similar in construction to the device of Fig. l,"the two coaxial lines 7 and '8, having respective inner'conductors 9 and 10, have their axes'perpendicular to the axis of'the guide and'are'arranged to act as coupling means to the guide, just asin Fig. 1. These lines 7 and 8 are at 90 with respect toeach other, as indicated, and are thus orthogonally positioned with re spect to each other. The lines 7 and 8 are at an angles with respect to therespective lines 2 and 3, as indicated in Fig. 5 for line 7. There is an 'axial spacing of one guide Wavelength between lines 2 and 3, and the same spacing between lines 7 and 8. For convenience and symmetry, the axial spacing between lines :3 and 7 is also ofthis same value,"though of course this is not necessary for proper operation of the device.

Forutilization of the diplexer of Fig.5 as a power combining system, the output of a "first transmitter (A) is applied by means of coaxial line 2 to the input circular Waveguide section 1, and the output of a second transmitter (B) is applied by means of coaxial line 3 to the input section 1. A power-'absorbing'resistor i1 is connected to coaxial line 7 of the output circular waveguide section 6, and a power-absorbing resistor 12 is "connected to coaxial line 8 of the output section-6. One or both of these resistors may represent a useful load such as a line extendin g to an antenna.

In operation, transmitter A with a 'power- P1 and current 11 is connectedto lineland transmitterB with a power P2 and current 12 isconnected toline 3, the two transmitters being assumed to be equal in frequency. As previously explained in connection with Fig. 1, the two transmitters are decoupled from each other because of the dominant mode (H11)- of the circular waveguide, and the orthogonal relation of the two lines 2 and 3. If the terminations at 7 and 8 match the guide, the current in resistor 11 is Squaring Equations 3 and 4, adding and simplifying,

we have or, since power is proportional to the square of the current,

It is possible to get a complete cancellation of current in one load and a complete addition in the other, for any ratio of input powers. For cancellation in resistor 12, I1 and I2 are in phase, so that I. P2 tan and tan 0- (7) Substituting these values in Equation 3, we have sin 0 I In I1 cos 6+I1 cos 6 003 0 (8) Substituting the values of Equation 7 in Equation 4, we have From the above analysis, it may be seen that when the circular waveguide diplexer of Fig. 5 is used as a powercombining system, it is readily possible to obtain complete power cancellation in one power-absorbing resistor (e. g., resistor 12), all the power input from both transmitters A and B) then appearing in the other powerabsorbing resistor (e. g., resistor 11), without any interaction or coupling between the two transmitters themselves. In the case where P1=P2, 0 should be 45.

In practice, lines 2 and 3 will not be in the same transverse plane, in order to get better decoupling, and likewise for lines 7 and 8 (they may, in fact, be separated a distance of one guide wavelength, as previously described). This does not change the operation of the device. The foregoing analysis has assumed a separation of one guide Wavelength. However, if the coupling means (coaxial lines) 2 and 3 are a half-guide-wavelength apart the conditions for cancellation in resistor 12 are the same as above except that I and 12 are 180 out of phase.

The principle underlying the use of the double-section circular waveguide diplexer as a power-combining systerm may be explained in another way. If a power nW is fed into arm 2 and W into arm 3, the waves being assumed in phase, the electric field strengths produced in the circular guide are proportional to x/nW and /W. The resultant of the two field vectors has a magnitude of 1 a. plane of polarization=6:13am- 11.

If the orientation of arm 7 (with its resistor 11) is in the direction of the resultant and that of arm 8 (with its resistor 12) is perpendicular to it, the total power, (n+1)W, leaves the system by arm 7, with arm 8 receiving no power.

In case the transmitters A and B being diplexed have different frequencies, for example if they are the picture and sound (video and audio) transmitters of a television broadcasting station, it is not possible to get a complete cancellation in either of the loads 11 or 12. In this case 0 would be 45 and both lines or arms 7 and 8 would be connected to a transmitting antenna, such as a turnstile antenna, which requires a double feed. Then, in conuection with Fig. 5, the resistors 11 and 12 would both draw power from the two transmitters, and these would represent the two respective arms of a turnstile antenna.

' The device described in connection with Fig. 5 is capable of handling very high power.

Reference will now be made to Fig. 6, which discloses the double-section circular waveguide diplexer of Fig. 5 used for operating two injection-locked magnetrons in parallel, for supplying power to a single common load. It will be assumed that the two magnetrons are of equal power. Since in Fig. 5 the axial spacing between the adjacent arms 2, 3, 7 and 8 is one guide wavelength, as previously described, all of the side arms may be considered to be located in the same transverse plane, and they are so represented in Fig. 6.

A magnetron M is connected to arm (coaxial transmission line) 2 at a distance I from the junction of the circular guide 1 and this line, so that arm 2 has a length I. Magnetron M1 is of conventional design, as indicated, and comprises a centrally-located cathode 13 surrounded by a plurality of radially-directed anode vanes 14 the spaces between which form cavity resonators in which energy of microwave frequency is developed when the magnetron is properly energized and supplied with the required magnetic field. Wave energy produced by magnetron oscillator M1 is applied to line 2 through the agency of a coupling loop 15 which is joined to the inner conductor 4 of line 2 and is positioned within one of the cavity resonators of magnetron M1.

The magnetron oscillator M1 may be aplitude modulated, for example by a television video signal, in any suitable manner (not shown), for example by modulation of the anode voltage thereof. Amplitude modulated wave energy is then supplied by magnetron M1 to the circular waveguide 1.

A similar magnetron M2 is connected to arm (coaxial transmission line) 3 at a diiferent distance from the junction of the circular guide 1 and this line, so that arm 3 has a length where )t is the free-space wavelength at the frequency of operation. Magnetron M2 is of the same design as magnetron M1, and functions similarly to develop energy of microwave frequency when properly energized. This energy is applied to line 3 by a coupling loop 15.

Likewise, the magnetron oscillator M2 may be amplitude modulated in any suitable manner, for example by modulation of the anode voltage thereof. Amplitude modulated wave energy is then supplied by magnetron M to the circular waveguide 1.

In Fig. 6, the angle 0 (between arms 2 and 7, or between arms 3 and 8) is made equal to 45 as indicated, since the magnetron powers are assumed to be equal. An injection power source 16, illustrated as a magnetron, is connected to arm (coaxial transmission line) 7. This source is used for frequency locking or stabilization of the two main magnetrons M1 and M2 by the injection locking process and therefore must itself have a very stable frequency. If this source 16 is a magnetron, it must be frequency-stabilized in any suitable manner (not shown), such as through the agency of a crystal oscillator. The injection power source 16 need be of only very small output power capacity as compared to the output powers of the magnetrons M1 and M2, and in fact source 16 needs to develop only a small fraction of the power developed by either one of the magnetrons M1 or M2. The injection power source 16 produces unmodulated CW energy, preferably of a frequency equal to the frequency of operation of magnetrons M1 and M2, and this energy is coupled to guide 1 by means of line 3. It should be understood that, as in Fig. 5, arms 2 and 3 are near one end of the waveguide while arms 7 and 8 are near the other end of the waveguide.

each other because of the dominant mode (H11) The single common load, illustrated as resistor lz, is connected to arm (coaxial transmission line) .8,-just as in Fig. 5. As in Fig. 5, the two lines 2 and '3-of'the finput section are orthogonally related'toseachother, 'thatis, they are at 90 relative to each other, and the two lines 7 and 8 of the output section are orthogonally related to each other. As previously explained in connection with Fig. l, the two magnetrons M1 and M2 are'decoupled from of the circular waveguide, and the orthogonal relation of the two lines 2 and 3.

If the injection source 16 feeds a power W byway of arm 7, the electric field strength created in .thecircular guide is proportional to ./W and the plane of polarization is inclined at 45 with the horizontal direction. After resolving this vector into horizontal and vertical components, it may be seen that the locking signal coupled into each of arms 2 and 3 is equal to These two waves are equal in magnitude and .are in time phase at .the respective junction points of the .two arms with the circular guide. Let [8 be the phase constant of the side-arm coaxial lines. Magnetron M1 is therefore locked by a signal proportional to while magnetron M2 is locked by a signal proportional to since magnetron M2 is at a distance a quarter-wavelength farther away from its junction point than is magnetron M1. The load arm 8 receives'no locking power (since the electric vector representing the locking Wave is at right angles to arm 8 and has no component parallel to the longitudinal axis of said arm), thus preventing any reduction of the depth of modulation at the load.

The locking power, injected into the magnetrons M1 and M2, serves to phase-lock their outputs with that of the stable frequency injection source 16, thus stabilizing the frequencies and phases of the two main magnetrons, this locking occurring as described in the aforementioned Koros application, Serial No. 177,455. Since the locking signal applied to magnetron M2 has a phase lag of 90 from that applied to magnetron M1 (due to the quarter-wave additional length in line 3 as compared to line 2), the output of magnetron M2 has a phase lag (at its output coupling loop), with respect to magnetron M1, of 90.

Under the assumption that the magnetrons M1 and M2 are locked in phase with the locking or synchronizing waves, the generated voltages of M1 and M2 in the cir cular waveguide may be written, respectively, as

since the magnetron M2 has a locked-phase lag of 90 and since the output of magnetron M2 is-subjected to an additional phase delay of 90 in reaching the circular guide, due to the quarter-wave" additional length of line 3 as compared to line 2. In other words, magnetron M2 excites the circular waveguide with a phase lag of 180 with respect to the excitation from magnetron M1. The polarization of the wave generated by M1 will, however, remain vertical and that of the wave generated by M2 will still be horizontal.

The vector E1 in Fig. 6 represents the power. of magnetron M1 in the circular guide and the vector E2 represents the power of magnetron M2 in'the' guide. DllBztO the 180 phase lag of the excitation'by magnetron M2 with respect to that by magnetron Mi, these vectors are in opposite directions with respect to their respective coaxial lines. The resultant Rofthe two vectors E1 and E2 possesses a polarization coincident with the axis of arm or line'8, but normal-to the axis of. arm or line 7. Thus, the magnetron powers are added inrthe load resistor 12 connected to line 8 and are cancelled insofar as any coupling to the line 7 or the injection source 16 is concerned. This is true since the electric vector R representing the resultant is parallel to thelongitudinal axis of line 8 but has no component parallel to the longitudinal axis of line 7, being perpendicular thereto. In other'words, with the addition of a quarter-wavelength in arm 3, the plane of polarization of'the resultant generated Waves is caused to shift by from'that of the original locking wave in arm 7, so that the 'magnetrompowers are coupled entirely to the load in arm 8 and the injection source 16 is isolated from the main magnetron power.

If the magnetrons M1 and M2 have unequal powers the angle 6 between arms 2 and 7 can be chosen so that the magnetron power fed to the injection source is zero. This may be done by causing the polarization plane assumed by the resultant field in the circular guide to lie parallel to the load arm 8 and normal to the injection arm 7. If the main magnetron powers are unequal, however, more injection power is required to get the same ratio of injection power to magnetron power in the higher power magnetron. The lower'power magnetron then has more injectionpowerthan needed. The optimum performance is obtained-when the two magnetrons have equal power.

The device of'Fig. 6, being like that of Fig. 5, is capable of handling very high power.

It may be'seen, from the foregoing, that successful operation of the circular waveguide diplexer depends on a definite andproper orientation of the polarization plane which is assumed by the. resultant of the fields generated by the two magnetrons. The resultant field in the circular guide should'be plane-polarized in a direction parallel to the load arm 8 and normal to the injection arm 7. To obtain thiscondition, the two magnetron generated waves should be equal in magnitude, exactly 90 in space phase, and exactly in time quadrature. Therefore, there are several requirements which must be met by the physical structure of the diplexer. First, the axial spacing between the two magnetron arms 2 and 3 should be :an exact multiple of a half guide wavelength. If this relation is not maintained, the two magnetron voltages in the circular guide at any plane between arms 3 and 8 will have a time phase difference and their resultant will be a circularly or an elliptically polarized wave. Next, the angular spacingbetween arms 2 and 3, and that between arms 7 and 8, should be exactly 90. A departure from this azimuth gives rise to imperfect isolation between the two side arms in the input as Well as the output section; also, the two magnetron fields will not maintain exactly a space phase quadrature. Next, the angular spacing between arms 2 and 7 should v be 45". This entails that the voltages generated bythe two magnetrons ought to have the same magnitude. Finally, the circular guide should have no eccentricity to cause mode instability.

In addition to the constraints put upon thephysical structureof the diplexer, proper adjustments should be made in the operating condition of the magnetrons, in order to promote their oscillation in equal amplitude, correct phase andrat the same frequency. First, the generated voltages of the two magnetrons should be of equal magnitude. If they are not, the angle between the plane of polarization of the resultant field and the y-axis will be difierent 1 from 45, in which event some power-from the magnetrons will be coupled to arm 7; this may cause damage-to the injection source 16. Next, thetwo magnetron voltages at the plane of arm 3 must be in time phaseto .give 'rise to a plane-polarized resultant wave. .When :these .two voltages have a phase difference, :thecombination-will result in a circularly or elliptically polarized wave, depending on the magnitude of this difierence in phase; in this case, the injection source 16 in arm 7 draws power. Next, the two magnetrons should oscillate at the same frequency, in order to produce a fixed polarization of the resultant wave. When magnetron M1 connected to arm 2 oscillates at a slightly different frequency from magnetron M2 connected to arm 3, the polarization plane of the resultant of the two magnetron fields will rotate on a plane perpendicular to the circular guide axis as time goes on; the resultant field vector varies both in magnitude and in direction from one instant to another and a weaving pattern will be produced. Then, the powers coupled into arms 7 and 8 will vary continuously with the passage of time.

In setting up the locking of two magnetrons as disclosed in Fig. 6, it is imperative that the magnetrons and the injection source be pretuned to oscillate at approximately the same frequency. With the injection source 16 first set in operation, the magnetrons M1 and M2 are next switched on to effect the locking action. Any appreciable disparity in their free-running frequencies would make it diflicult for the locking source to lock the magnetrons in step, on the one hand, and would provide a resultant field vector (the resultant of the fields produced by the two magnetrons) such as to couple energy to the injection arm 7, on the other hand.

What is claimed is:

1. A power-combining system comprising a length of circular waveguide capable of propagating therein wave energy in the dominant mode, a pair of orthogonallyrelated energy coupling members arranged in energyinterchanging relationship to said guide, first and second magnetron oscillators, means coupling the output of said first oscillator to one of said members, means coupling the output of said second oscillator to the other of said members, two arms coupled to said guide in energyinterchanging relationship thereto, the planes of polarization of wave energy propagated in said two arms lying at right angles to each other, means coupling the output of a stable frequency source of injection power to one of said arms, and means coupling a load to the other of said arms.

2. A power-combining system comprising a length of circular waveguide capable of propagating therein wave energy in the dominant mode, a pair of orthogonallyrelated coaxial lines arranged perpendicularly to the longitudinal axis of said guide and coupled in energyinterchanging relationship to said guide, first and second magnetron oscillators, means coupling the output of said first oscillator to one of said lines, means coupling the output of said second oscillator to the other of said lines, the lengths of said lines, between the respective magnetron outputs and the guide, diifering by an odd multiple of a quarter-wavelength at the operating frequency of the magnetrons, two arms coupled to said guide in energy-interchanging relationship thereto, the planes of polarization of wave energy propagated in said two arms lying at right angles to each other, means coupling the output of a stable frequency source of injection power to one of said arms, and means coupling a load to the other of said arms.

3. A power-combining system comprising a length of circular waveguide capable of propagating therein wave energy in the dominant mode, a pair of orthogonallyrelated energy coupling members arranged in energyinterchanging relationship to said guide, first and second magnetron oscillators operating at the same frequency, means coupling the output of said first oscillator to one of said members, means coupling the output of said second oscillator to the other of said members, two arms coupled to said guide in energy-interchanging relationship thereto, the planes of polarization of wave energy propagated in said two arms lying at right angles to each other, means coupling the output of a stable frequency 10 v source of injection power operating at the frequency of said oscillators to one of said'arms, and means coupling a load to the other of said arms.

4. A power-combining system comprising a length of circular waveguide capable of propagating therein wave energy in the dominant mode, a pair of orthogonallyrelated coaxial lines axially spaced from each other along said guide a distance of a multiple of a half-wavelength at the operating frequency and coupled in energy-interchanging relationship to said guide, first and second magnetron oscillators, means coupling the output of said first oscillator to one of said lines, means coupling the output of said second oscillator to the other of said lines, two orthogonally-related energy coupling members arranged in energy-interchanging relationship to said guide, said members being angularly displaced from said pair of lines, means coupling the output of a stable frequency source of injection power to one of said members, and means coupling a load to the other of said members.

5. A power-combining system comprising a length of circular waveguide capable of propagating therein wave energy in the dominant mode, a pair of orthogonallyrelated coaxial lines coupled in energy-interchanging relationship to said guide, first and second magnetron oscillators, means coupling the output of said first oscillator to one of said lines, means coupling the output of said second oscillator to the other of said lines, two orthogonally-related coaxial lines coupled in energy-interchang ing relationship to said guide, said two last-named lines being angularly displaced from said first-named pair of lines, means coupling the output of a stable frequency source of injection power to one of said last-named lines, and means coupling a load to the other of said lastnamed lines.

6. A power-combining system comprising a length of circular waveguide capable of propagating therein wave energy in the dominant mode, a pair of othogonallyrelated coaxial lines coupled in energy-interchanging relationship to said guide, first and second magnetron oscillators, means coupling the output of said first oscillator to one of said lines, means coupling the output of said second oscillator to the other of said lines, the lengths of said lines, between the respective magnetron outputs and the guide, differing by an odd multiple of a quarterwavelength at the operating frequency of the magnetrons, two orthogonally-related coaxial lines coupled in energyinterchanging relationship to said guide, said two lastnamed lines being angularly displaced from said firstnamed pair of lines, means coupling the output of a stable frequency source of injection power to one of said lastnamed lines, and means coupling a load to the other of said last-named lines.

7. A power-combining system comprising a length of circular waveguide capable of propagating therein wave energy in the dominant mode, a pair of orthogonallyrelated coaxial lines coupled in energy-interchanging relationship to said guide, first and second magnetron oscillators operating at the same frequency, means coupling the output of said first oscillator to one of said lines, means coupling the output of said second oscillator to the other of said lines, the lengths of said lines, between the respective magnetron outputs and the guide, differing by an odd multiple of a quarter-wavelength at the operating frequency of the magnetrons, two orthogonallyrelated coaxial lines coupled in energy-interchanging relationship to said guide, said two last-named lines being angularly displaced from said first-named pair of lines, means coupling the output of a stable frequency source of injection power operating at the frequency of said oscillators to one of said last-named lines, and means coupling a load to the other of said last-named lines.

8. A power-combining system comprising a length of circular waveguide capable of propagating therein wave energy in the dominant mode, a pair of orthogonallyrelated coaxial lines axially spaced from each other along v1-1 said guide a distance of a multipleiof a ha1fwavelength at the operating frequency and coupled in energy-interchanging relationship ,to said guide, first and second magnetron oscillators, means coupling the output of said first oscillator to one of said lines, means coupling the output of said'secondtoscillator to the other of said lines, two orthogonally-related coaxial lines axially spaced from each other along said guide a distance of a multiple of a half-wavelength atisaid frequency and-coupled in energyinterchanging relationship to said guide, said two lastnamed lines being augularly displaced from said firstnamed; pair of lines and also'being axially displaced an appreciable distance fromsaid'first-tnamedpair of lines,

means coupling the output of a stable frequency source of injection power to one of said last-named lines, and means coupling a'load to the other of said'last-named lines.

References Cited inthe file of this patent UNITED STATES PATENTS

Images