Coupler having arbitary impedance transformation ratio and arbitary coupling ratio

Description

Background of the invention

1. Field of the invention

[0001] This invention is a circuit for power combining two or more amplifier elements at microwave frequencies employing a novel coupler which simultaneously couples with an arbitrary power coupling (combining or dividing) ratio and matches to an arbitrary impedance.

2. Description of the prior art

[0002] A prior art search was performed and uncovered the following U.S. patent references:

U.S. patent 3,772,616 is an impedance matching power divider. The power division ratio is always unity, unlike in the present invention. The patent speaks just in terms of power division, not power combining as in the present invention. The branches of the circuit are rectangular, not curved as in the present invention.

[0003] U.S. patent 3,516,024 is a coupler with no impedance transformation capability, and without the capabilities for an arbitrary power division or combining ratio.

[0004] U.S. patent 3,423,688 shows various couplers but does not show impedance transformation.

[0005] U.S. patent 3,654,570 is a hybrid junction device without having an arbitrary power coupling ratio. In order to fulfill the prophecy of column 3 line 53 et seq. that other impedance values can be accommodated, the patentee would have to use additional components not shown in his patent.

[0006] U.S. patent 4,127,831 is a coupler which does not match impedances other than 50 ohms. The function of the stubs is to broaden the 50 ohm bandwidth.

[0007] U.S. Patent 4,127,832 and 4,027,254 are couplers without an impedance transformation capability.

[0008] Secondary references are U.S. patents 3,327,130, 3,600,707, 4,016,503, and 3,063,026.

[0009] C. G. Montgomery et al "Principles of Microwave Circuits", 1948, First Edition, McGraw-Hill Book Company, Inc., New York (U.S.), pages 308-311, suggests on page 310 in relation to figure 9.38 that a directional coupler can be constructed wherein each branch line has a different characteristic impedance. It does not show how to construct such a coupler, nor does it suggest that it may be possible to calculate the values of characteristic impedances of each branch line in order to provide arbitrary power coupling and impedance matching ratios.

[0010] IEEE Transactions on Microwave Theory & Techniques, Vol. MTT-29, No. 4, April 1981 New York (U.S.) A. A. M. Saleh: "Planar Multiport Quadrature-Like Power-Dividers/Combiners" pages 332 to 337 discloses in figure 1 (b) the use of 3-d B quadrature hybrids for input-matched combining of two identical devices.

Summary of the invention

[0011] The present invention is a power coupler (divider or combiner) operable at microwave frequencies having both an arbitrary power coupling capability and an arbitrary impedance matching capability.

[0012] When used to power combine a pair of amplifier elements, the bandwidth of the resulting circuit is improved by an order of two to one over circuits of the prior art using separate devices for impedance transformation and power combining. The insertion loss is also significantly lower, and the size of the circuit is reduced.

[0013] The coupler is optimally a branch line coupler having four circularly curved arcuate branches, which provide a greater bandwidth than rectangular branches. The four branches are each an odd multiple of a quarter wavelength long at the center frequency. The output ports of the coupler are 90° out of phase.

[0014] Two of the branches of the coupler have normalized admittances equal to (Y/k)+Y where k is the desired power coupling ratio, and Y (the reciprocal of impedance ratio Z) is the desired ratio of load admittance to source admittance. The input branch has a normalized admittance equal to 1/VPand the output branch has a normalized admittance equal to Y/Vk.

[0015] These normalized admittances are then converted into actual admittances and then into appropriate heights and widths of the selected conductor. The branches each have a length equal to a quarter electrical wavelength at the desired center operating frequency.

Brief description of the drawings

[0016] These and other more detailed and specific objects and features of the present invention are more fully disclosed in the following specification, reference being had to the accompanying drawings, in which:

Figure 1 is a circuit diagram representing the state of the prior art;

Figure 2 is a circuit diagram of the impedance transforming couplers of the present invention in a power combining circuit of the present invention;

Figure 3 is a schematic model of the impedance transforming coupler of the present invention;

Figure 4 is a circuit which is equivalent to that depicted in Figure 3;

Figure 5 comprises two drawings, Figures 5a and 5b, which together are equivalent to the circuit depicted in Figure 4; and

Figure 6 is a shadow-graph of a branch line coupler constructed according to the teachings of the present invention.

Description of the preferred embodiments

[0017] The design of power combined solid state amplifiers capable of operation at microwave frequencies normally entails each amplifier element being fitted with impedance matching circuits on each end, so as to match the amplifier to the conventional 50 ohm source and 50 ohm load impedances. Symmetrical 50 ohm to 50 ohm couplers are then used to combine two or more amplifier elements to create the power combined solid state amplifier circuit. The power combining efficiency of the combined amplifier is determined by the impedance transformation circuit loss and the coupler circuit loss. The impedance transformation circuit loss is normally a direct function of the impedance transformation ratio. The higher the impedance transformation ratio, the higher the transformation circuit loss. The impedance transformation circuit loss may vary from 0.1 dB to as much as 0.5 dB, depending upon the transformation circuit used.

[0018] The power combining bandwidth of the circuit is limited by the impedance transformation ratio, the bandwidth of the couplers, and the bandwidths of the solid state amplifier elements. The bandwidth of each amplifier element is a strong function of the impedance transformation ratio also (the lower the impedance transformation ratio the wider the bandwidth). Therefore, ideally, the circuit loading impedance should be close to the inherent impedance level of each amplifier element for best bandwidth and minimum circuit loss.

[0019] Due to the relatively large active areas of power solid state devices, the input impedance and the output impedance are usually much less than the conventional 50 ohms which is employed throughout the transmission system. Extensive device evaluation of both FET and bipolar transistors from different manufacturers has shown that for most power transistors capable of several watts of power output, the optimum source and load impedances normally fall in the range between 1 and 20 ohms. Such low impedances are difficult to match to 50 ohm systems with large bandwidth. Also, the high impedance transformation ratio increases the circuit size as well as the circuit loss.

[0020] The present invention shows how to improve the circuit loss and size and the amplifier bandwidths of power combined amplifiers, by using impedance transformation couplers in the power combining of power amplifiers. These impedance transformation couplers are capable of matching arbitrary impedances and are simultaneously capable.of providing arbitrary power coupling ratios. These couplers can be used as power dividers and as power combiners with equal facility. With devices having source and load impedances less than 50 ohms, each device is connected directly to the low impedance portion of each coupler. Using the approach of the present invention, the bandwidth of the amplifier circuit can be improved by an order of two to one over the prior art, and the matching circuit loss can be reduced. In addition, the individual line or branch impedances of the impedance transformation coupler are lower than for the 50 ohm to 50 ohm couplers used in the prior art, resulting in a further improvement of the loss characteristics of the power combining circuit.

[0021] The problems with the prior art are graphically illustrated in Figure 1, which shows the power combining of two field effect transistors,11 and 12, each having an optimum source impedance of 10 ohms and an optimum load impedance of 20 ohms. Other types of amplifier elements could be substituted for purposes of this discussion. The gate of FET 11 is connected through impedance transformation circuit 13 to port 2 of coupler 17. The gate of FET 12 is connected through impedance transformation circuit 15 to port 3 of coupler 17. The sources of each of FET's 11 and 12 are grounded. The drain of FET 11 is connected through impedance transformation circuit 14 to port 1 of coupler 18. The drain of FET 12 is connected through impedance transformation 16 to port 4 of coupler 18. The input signal is applied to port 1 of coupler 17. Port 4 of coupler 17 is connected via impedance 19 to ground. The output signal appears at port 3 of coupler 18. Port 2 of coupler 18 is connected via impedance 20 to ground. The values of impedances 19 and 20 are typically each 50 ohms, representing the characteristic impedance of the system. The function of impedance transformation circuits 13,14,15, and 16 is to match the optimum input and output impedances of each of the individual devices 11 and 12 to 50 ohms. Where FET's having the specified optimum impedances were tested, the circuit loss introduced by these impedance transforming elements was 0.4 dB. The couplers 17 and 18 added an additional loss of 0.15 dB, which made a total RF circuit loss of 0.55 dB.

[0022] Figure 2 shows how the present invention remedies the defects of the prior art. In Figure 2 the impedance transformation is performed simultaneously with the power coupling function. The gate of FET 11 is connected directly to port 2 of coupler 27. The gate of FET 12 is connected directly to port 3 of coupler 27. The drain of FET 11 is connected directly to port 1 of coupler 28. The drain of FET 12 is connected directly to port 4 of coupler 28. Other connections are identical to those in Figure 1. Since there are no separate impedance transformations to be performed, the size of the circuit is kept to a minimum. The loss from each coupler is less than 0.1 dB, since the impedances of the branches of the coupler are kept to lower values than before (15.81 ohms and 22.36 ohms for the cross-branches versus 35 ohms of the prior art). The total circuit loss is therefore, .20 dB, which is 0.35 dB better than the conventional approach.

[0023] While the above Figure illustrates the case where a power combining and power dividing ratio of 1 was desired, the following analysis shows how one may construct a coupler having both arbitrary impedance transforming ratio and arbitrary power coupling ratio.

[0024] The general form of the coupler of the present invention is depicted in Figure 3. This is a model of an asymmetric coupler having arbitrary coupling (i.e., dividing or combining) ratio and arbitrary impedance matching capability. The. circuit comprises four ports, designated as ports 1, 2, 3, and 4 respectively, starting at the upper left and proceeding in a clockwise direction. Ports 1 and 4 are input ports; ports 2 and 3 are output ports. Signals applied at port 1 can be reversed with signals applied at port 4, and signals applied at port 2 can be reversed with signals applied at port 3 because the coupler is symmetric about a horizontal line bifurcating the a and c normalized admittances V1, V2, V3, and V4 are the voltage ratios (i.e., the voltage divided by one volt to make a unitless value), at ports 1, 2, 3, and 4 respectively.

[0025] Port 1 is connected to port 2 via a branch having normalized admittance b. Port 4 is connected to port 3 via a second branch having normalized admittance b. Port 1 is connected to port 4 via a third branch having normalized admittance a. Port 2 is connected to port 3 via a fourth branch having normalized admittance c.

[0026] 81, 82, 83, and 84 are impedances connected between ports 1, 2, 3, and 4, respectively, and ground. Impedances 81 and 84 are normally equal to each other and impedances 82 and 83 are normally equal to each other. The source impedance Z81=Z84, and the load impedance is Z82=Z83, when Zn is the impedance of component n.

[0027] The following derivation describes how one obtaines values for a, b, and c as a function of k, the desired power coupling ratio of the coupler, and Y, the desired admittance transformation ratio of the coupler.

[0028] Assume that an input signal having a value of 1 (e.g., 1 volt) is applied at port 1. This is equivalent to a signal of applied to port 1, plus another signal having a value applied to port 1, plus a signal having a value -1/2 applied to port 4, plus a signal having a value of +2 applied to port 4 (see Figure 4).

[0029] Figure 5 depicts the equivalency of this to the sum of two circuits. The first circuit, Figure 5a, has a signal of applied at port 1, and a signal of 2 applied at port 4. The second circuit, Figure 5b, has a signal of 2 applied at port 1 and a signal of -1/2 applied at port 4. The circuit of Figure 5a is the same as that of Figure 4 except that an imaginary horizontal line has been drawn separating normalized admittances a and c such that the impedance Z' is infinite and the admittance Y' is zero along this line. This line divides the circuit into an upper portion 91 and an equivalent lower portion 92.

[0030] Similarly, the circuit of Figure 5b is equivalent to the circuit of Figure 4 except that an imaginary horizontal line has been drawn severing normalized admittances a and c such that the impedance Z' is zero and the admittance Y' is infinite along this line. This line divides the circuit into an upper portion 93 and an equivalent lower portion 94 as shown in the drawing.

[0031] R1 is the voltage ratio reflected back into an input port of Figure 5a for a signal having value 1, so sR1 is reflected back into each of ports 1 and 4 of Figure 5a. T1 is the voltage ratio transmitted into an output port of Figure 5a for a signal having a value 1, so 1/2 T1 is transmitted into each of ports 2 and 3 of Figure 5a. R2 is the voltage ratio reflected into an input port of Figure 5b for a signal having value 1, so sR2 is reflected back into port 1 of Figure 5b and -ZR2 is reflected back into port 4 of Figure 5b. T2 is the voltage ratio transmitted into an output port of Figure 5b for a signal having a value 1, so 1/2 T2 is transmitted into port 2 of Figure 5b and -ZT2 is transmitted into port 3 of Figure 5b. Thus, it is seen that:

[0032] Let L equal the electrical length of each of the four branch lines a, b, c, d in Figure 4. Thus, in Figure 5a and Figure 5b the lengths of normalized admittances b are each L, and the length of the remaining stubs of normalized admittances a and c are U2 in each case. A is the wavelength at the central operating frequency. Let M1 be the ABCD matrix of circuit 91. Then:

[0033] Let M2 be the ABCD matrix of circuit 93. Then:

[0034] A good length for each conductor of a branch line coupler is a quarter wavelength. Thus, let L=N4. Then:

[0035] 

[0036] Similarly,

[0037] Now, the definition of an ABCD matrix yields, for circuit 91:

[0038] where Ein is the input voltage, lin is the input current, Eout is the output voltage, and lout is the output current. Thus,

and

[0039] where Yout, the output admittance, is equal to lout/Eout. If we normalize the source impedance against the characteristic impedance of the system, then the source impedance is one, and the output admittance Yout equals Y, the desired admittance transformation ratio of the coupler. The input impedance is Zin.

[0040] According to the definition of reflected voltage ratio from transmission line theory.

[0041] The definition of transmitted voltage ratio T1 is:

[0042] Similarly,

and

[0043] For a perfect match and perfect isolation, V1 equals V4 equals zero. Thus, R1 equals zero and from equation 12:

[0044] Now, from equations 2 and 3,

[0045] For an ideal coupler, the power appearing at port 4 is zero because all the input power appears at ports 2 and 3. Thus, for a normalized input power of 1:

[0046] The desired power coupling ratio of the coupler is k, the ratio of the power at port 2 to the power at port 3.

Thus,

[0047] Substituting from equations 19 and 20:

[0048] We have thus specified the normalized admittances a, b, and c as a function of the desired arbitrary power division ratio, k, and the desired arbitrary admittance transformation ratio, Y. It is now a straightforward task to convert these normalized admittances into actual admittances by the formula "actual admittance=(normalized admittance) (source admittance)" and then into physical dimensions for certain conductors. Bahl and Trivedi, "A Designer's Guide to Microstrip Line," Microwaves, May, 1977, p. 174 et seq. A circuit thus built will work extremely well over a wide range of frequencies. At the center frequency, the phase differential between ports 2 and 3 is exactly 90° and the voltage standing wave ratio of the coupler is exactly 1 to 1. These parameters deviate but slightly as the frequency is moved away from the center frequency.

[0049] To make the coupler as broadbanded as possible it is desirable to curve into a circular arc each of the four conductors comprising the four branches of the coupler. This is depicted in Figure 6, which is a shadowgraph tracing of a C-band hybrid (i.e., k≈1) coupler having a 50 ohm output load impedance and a 16 ohm input impedance. Isolation between the two output ports (ports 2 and 3) was measured at better. than 28 dB. The coupling variation between the two output ports was less than 0.13 dB and the mid-band insertion loss was 0.10 dB. The measured performance of the coupler corresponded closely to the theoretical calculation. A complete power combined FET amplifier was designed and fabricated using the coupler depicted in Figure 6.

[0050] Comparing the test results to those from the conventional approach showed that the impedance transformation coupler significantly improved the amplifier bandwidth and the amplifier power output.

Claims

1. A quadrature directional coupler comprising first (a), second (c), third (b), and fourth (b) transmission line branches having precalculated admittances building a planar network, said coupler having any desired impedance transformation capability; wherein the first branch is connected to each of the third and fourth branches, the second branch is connected to each of the third and fourth branches, the third branch is connected to each of the first and second branches, and the fourth branch is connected to each of the first and second branches; wherein a first port (1) is at the intersection of said first and third branches, a second port (2) is at the intersection of the third and second branches, a third port (3) is at the intersection of the second and fourth branches and a fourth port (4) is at the intersection of the first and fourth branches; wherein said first branch was normalized admittance equal to 1/Vk;

said second branch has a normalized admittance equal to 1/(ZVk); and

said third and fourth branches each have normalized admittances equal to 1/1/!Zk)+(1/Z);

where Z is the desired impedance transformation ratio of the coupler and k is the desired power coupling ratio, whereby K is not equal to one.

2. The coupler of claim 1 wherein said coupler can be used as both a power divider and a power combiner.

3. The coupler of claim 1 wherein said coupler is operable at microwave frequencies.

4. The coupler of claim 1 wherein the length of each branch is equal to an odd multiple of one quarter of the wavelength at the central operating frequency of the coupler.

5. The coupler of claim 1 wherein the branches of said coupler are curved into circular arcs.

6. A microstrip circuit operable at microwave frequencies for combining the signal power of two amplifiers in which the substantially identical optimum input impedances of said amplifiers differ from the input impedance of said circuit, said circuit comprising: a quadrature directional coupler according to claim 1, the coupler being a power dividing coupler consisting essentially of four series-connected microstrip branches and having two input ports and two output ports; and the circuit further comprising:

two amplifiers, the input of each of which is connected to a different one of said output ports of said power dividing coupler; and

a power combining coupler having two input ports and two output ports; wherein

the outputs of each of said amplifiers are connected to a different one of said input ports of said power combining coupler;

wherein a different amount of power flows through each of said amplifiers; and

said power dividing coupler has a power division ratio other than one and simultaneously matches the input impedance of the circuit to the optimum input impedances of said amplifiers (Figure 2).

7. A microstrip circuit operable at microwave frequencies for combining the signal power of two amplifiers in which the substantially identical optimum output impedances of said amplifiers differ from the output impedance of said circuit, said circuit comprising:

a power dividing coupler having two input ports and two output ports;

two amplifiers, the input of each of which is connected to a different one of said output ports of said power dividing coupler; and

a quadrature directional coupler according to claim 1, the coupler being

a power combining coupler consisting essentially of four series-connected microstrip branches and having two input ports and two output ports;

wherein the outputs of each of said amplifiers are connected to a different one of said input ports of said power combining coupler;

wherein a different amount of power flows through each of said amplifiers; and

said power combining coupler has a power combination ratio other than one and simultaneously matches the output impedance of the circuit to the optimum output impedances of said amplifiers (Figure 2).

Ansprüche

1. Richtungs-Quadratur-Koppler bestehend aus einem ersten (a), zweiten (c), dritten (b) und vierten (b) Übertragungsleitungszweig mit vorkalkulierter Admittanz, die ein ebenes Netzwerk bilden wobei der Koppler eine beliebige gewünschte Impendanztransformationsmöglichkeit hat, bei dem der erste Zweig mit dem dritten und vierten Zweig verbunden ist, der zweite Zweig mit dem dritten und vierten Zweig verbunden ist, der dritte Zweig mit dem ersten und zweiten Zweig verbunden ist und der vierte Zweig mit dem ersten und zweiten Zweig verbunden ist;

bei dem ein erstes Port (1) sich am Verbindungspunkt vom ersten und dritten Zweig befindet, ein zweites Port (2) sich am Verbindungspunkt des zweiten und dritten Zweiges befindet, ein drittes Port (3) sich am Verbindungspunkt des zweiten und vierten Zweiges befindet und ein viertes Port (4) sich am Verbindungspunkt des ersten und vierten Zweiges befindet;

bei dem der ersten Zweig eine normierte Admittanz gleich 1/Vk hat;

der zweite Zweig eine normierte Admittanz gleich 1/(ZVk); und

der dritte und vierte Zweig jeder eine normierte Admittanz gleich Vt/(Zk)+(1/Z);

wobie Z das gewünschte Impedanztransformationsverhältnis des Kopplers ist und k das gewünschte Leistungskopplungsverhältnis ist, so daß K nicht gleich eins ist.

2. Der Koppler nach Anspruch 1 bei dem der Koppler sowohl als Leistungsteiler als auch als Leistungskombinationseinrichtung verwendet werden kann.

3. Der Koppler nach Anspruch 1 bei dem der Koppler für Mikrowellenfrequenzbetrieb geeignet ist.

4. Der Koppler nach Anspruch 1 bei dem die Länge jedes Zweiges gleich einem ungeraden Vielfachen von einer Viertelwellenlänge bei der Betriebsmittenfrequenz des Kopplers ist.

5. Der Koppler nach Anspruch 1 bei dem die Zweige des Kopplers in Kreisbogen gekrümmt sind.

6. Mikrostrip-Schaltung zum Betrieb bei Mikrowellenfrequenzen zum Kombinieren der Signalleistung von zwei Verstärkern, bei dem die im wesentlichen identischen optimalen Eingangsimpedanzen der Verstärker sich von der Eingangs impedanz der Schaltung unterscheiden, wobei diese Schaltung besteht aus:

einem Richtungs-Quadratur-Koppler nach Anspruch 1, bei dem es sich um einen Leistungsteilerkoppler handelt, der im wesentlichen aus vier in Reihe geschalteten Mikrostrip-Zweigen besteht und zwei Eingangsports und zwei Ausgangsports hat; wobei die Schaltung weiter besteht aus:

zwei Verstärkern, von denen jeweils der Eingang an einen anderen der Ausgangsports des Leistungsteilerkopplers angeschlossen ist; und

einem Leistungs-Kombinations-Koppler mit zwei Eingangsports und zwei Ausgangsports; bei dem

die Ausgänge jedes der Verstärker mit einem anderen der Eingangsports des Leistungs-Kombinations-Kopplers verbunden sind;

bei dem die unterschiedlicher Leistungsbetrag durch jeden der Verstärker fließt; und

der Leistungsteilerkoppler ein anderes Leistungsteilungsverhältnis als eins hat und gleichzeitig die Eingangsimpedanz der Schaltung an die optimalen Eingangsimpedanzen der Verstärker anpaßt (Figur 2).

7. Eine Mikrostrip-Schaltung für Betrieb bei Mikrowellenfrequenzen zum Kombinieren der Signalleistung von zwei Verstärkern, bei dem die im wesentlichen identischen optimalen Ausgangsimpedanzen der Verstärker sich von der Ausgangsimpedanz der Schaltung unterscheiden, wobei die Schaltung besteht aus:

einem Leistungsteilerkoppler mit zwei Eingangsports und zwei Ausgangsports;

zwei Verstärkern, von denen jeweils der Eingang an eine andere der Ausgangsports des Leistungsteilerkopplers angeschlossen ist; und

einen Richtungs-Quadratur-Koppler nach Anspruch 1, wobei der Koppler ein Leistungs-Kombinations-Koppler ist, der im wesentlichen aus vier in Reihe geschalteten Mikrostrip-Zweigen besteht und zwei Eingangsports und zwei Ausgangsports hat;

bei dem die Ausgänge jedes der Verstärker an einen anderen der Eingangsports des Leistungs-Kombinations-Kopplers angeschlossen sind;

bei dem ein unterschiedlicher Leistungsbetrag durch jeden der Verstärker fließt; und

der Leistungs-Kombinations-Koppler ein anderes Leistungskombinationsverhältnis als eins hat und gleichzeitig die Ausgangsimpedanz der Schaltung an die optimalen Ausgangsimpedanzen der Verstärker anpaßt (Figur 2).

Revendications

1. Coupleur directionnel un quadrature comprenant une première (a), une seconde (c), une troisème (b) et une quatrième (b) lignes de transmission en dérivation, ayant des admittances calculées préalablement et constituant un réseau plan, le coupleur pouvant donner toute transformation voulue d'impédance, dans lequel la première branche est connectée à chacune des troisième et quatrième branches, la seconde branche est connectée à chacune des troisième et quatrième branches, la troisième branche est connectée à chacune des première et seconde branches, et la quatrième branche est connectée à chacune des première et seconde branches,

dans lequel une première voie (1) se trouve à l'intersection de la première et de la troisième branches, une seconde voie (2) se trouve à l'intersection de la troisième et de la seconde branches, une troisième voie (3) se trouve à l'intersection de la seconde et de la quatrième branche et une quatrième voie (4) se trouve à l'intersection de la première et de la quatrième branche,

dans lequel la première branche a une admittance normalisée égale à 1/Vk,

la seconde branche a une admittance normalisée égale à 1/(ZVk), et

la troisième et la quatrième branches ont chacune des admittances normalisées égales à

Z étant le rapport voulù de transformation d'impédance du coupleur et k étant le rapport voulu de couplage d'énergie, k n'étant pas égal à 1.

2. Coupleur selon la revendication 1, dans lequel le coupleur peut être utilisé à la fois comme diviseur d'énergie et comme combinateur d'énergie.

3. Coupler selon la revendication 1, dans lequel le coupleur peut fonctionner à des hyperfréquences.

4. Coupleur selon la revendication 1, dans lequel la longueur de chaque branche est égale à un multiple impair du quart de la longueur d'onde à la fréquence centrale de fonctionnement du coupleur.

5. Coupleur selon la revendication 1, dans lequel les branches du coupleur sont courbées en arc de cercle.

6. Circuit à microlignes plates destiné à fonctionner en hyperfréquence et destiné à combiner l'énergie des signaux de deux amplificateurs, dans lequel les impédances optimales sensiblement identiques d'entrée des amplificateurs diffèrent de l'impédance d'entrée du circuit, le circuit comprenant

un coupleur directionnel en quadrature selon la revendication 1, le coupleur étant

un coupleur de division d'énergie constitué essentiellement de quatre branches à microlignes plates montées en série et ayant deux voies d'entrée et deux voies de sortie, et le circuit comprenant en outre

deux amplificateurs, l'entrée de chacun d'eux étant connectée à une voie différente de sortie du coupleur de division d'énergie, et

un coupleur de combinaison d'énergie ayant deux voies d'entrée et deux voies de sortie, dans lequel

les sorties des amplificateurs sont reliées à des voies différentes d'entrée du coupleur de combinaison d'énergie,

dans lequel des quantités différentes d'énergie circulent dans les amplificateurs, et

le coupleur de division d'énergie a un rapport de division d'énergie différent de 1 et assure simultanément l'adaptation de l'impédance d'entrée du circuit aux impédances optimales d'entrée des amplificateurs (figure 2).

7. Circuit à microlignes plates, destiné à fonctionner en hyperfréquence afin qu'il combine l'énergie des signaux de deux amplificateurs, les impédances optimales sensiblement identiques de sortie des amplificateurs différant de l'impédance de sortie du circuit, le circuit comprenant un coupleur de division d'énergie ayant deux voies d'entrée et deux voies de sortie,

deux amplificateurs, l'entrée de chaque amplificateur étant reliée à une voie différente de sortie du coupleur de division d'énergie, et

un coupleur directionnel en quadrature selon la revendication 1, le coupleur étant

un coupleur de combinaison d'énergie formé essentiellement de quatre branches de microlignes plates connectées en série et ayant deux voies d'entrée et deux voies de sortie,

dans lequel les sorties des amplificateurs sont connectées à des voies différentes d'entrée du coupleur de combinaison d'énergie,

dans lequel des quantités différentes d'énergie circulent dans les amplificateurs, et

le coupleur de combinaison d'énergie a un rapport de combinaison d'énergie différent de 1 et assure simultanément l'adaptation de l'impédance de sortie du circuit aux impédances optimales de sortie des amplificateurs (figure 2).

Drawing