QUASI-OPTICAL WAVEGUIDE BEAMFORMER WITH SUPERIOR PARALLEL PLATES

Description

Domaine technique

[0001] The invention relates generally to the field of telecommunications, and in particular to quasi-optical beamformers (QOBs) for active multi-beam antennas.

[0002] Quasi-optical beamformers can be carried on satellites or ground stations. Antennas using such beamformers can operate in transmission or reception mode, reciprocally.

[0003] A quasi-optical beamformer is a focusing (receiving) and collimating (transmitting) device. The figure 1 represents a state-of-the-art quasi-optical beamformer that can be applied, for example, to pillbox, continuous delay lens or Rotman beamformers. A quasi-optical beamformer typically incorporates a parallel plate guide 16, connecting beam ports 17 and network ports 18. The parallel plate waveguide 16 allows waves to be guided in TEM mode (acronym for “Transverse Electric Magnetic”), in which the electric field E and the magnetic field H evolve in directions perpendicular to the propagation direction X.

[0004] The wavefronts are curved in the XY plane. In order to compensate for the curvature of the wavefront, a quasi-optical device 23 is introduced between the beam ports and the array ports. This quasi-optical device may be, for example, a lens as used for continuous delay lenses or a reflector as used for pillbox beamformers. Each array port 18 is connected to an amplifier 19 followed by a radiating element 20 via a delay line 21 and an amplifier port 22. It transforms the cylindrical waves emanating from the beam ports into plane waves radiated by the radiating panel of the active multibeam antenna.

[0005] Quasi-optical beamformers produce multiple axially aligned beams, which usually intersect at a gain level that can be up to 10 dB lower than the maximum beam gain, as illustrated by figure 2 Such limitations are classic and usually observed for any multi-beam antenna associating an optical system (for example a reflector, a lens) and a focal network of passive multi-sources, each of them defining a spot access.

[0006] This level of beam overlap comes from a compromise on the size of these sources which must meet two opposing constraints: on the one hand, they must be wide enough to properly illuminate the optical system, and thus avoid losses through overflow, and on the other hand, they must be close enough for the beams to overlap.

[0007] When a geographical area is covered by an antenna producing this beam grouping, some ground stations are then exposed to an antenna gain reduced by these overlapping losses. It is therefore desirable to minimize these overlapping losses, and therefore to produce multiple beams which overlap at a high gain level.

[0008] Several solutions have been considered to minimize losses related to beam crossing.

[0009] For example, it is known to use two quasi-optical beamformers with interleaved sources, as disclosed for example in patent application WO 2013 / 110793 A1. The use of these two formers makes it possible to double the beam density over a given angular sector. However, this solution requires two quasi-optical beamformers, as well as a combining stage. This results in a greater mass, as well as a very significant increase in complexity for the case of two-dimensional beamforming.

[0010] Other solutions use smaller sources with sliding recombination of two adjacent sources, as disclosed in the article "A theoretical limitation on the formation of lossless multiple beam antennas" (J. L. Allen, IRE Trans., 1961, AP-9, pp. 350-352). This approach allows for the production of equivalent sources that are sufficiently large and partially overlap so that the associated beams will intersect at a higher level. However, this solution requires the addition of a circuit combining a divider and a combiner, which complicates the formatter and generates additional losses.

[0011] In other solutions, apodization of the signal at the output ports is used to broaden the main lobe of each beam while lowering the level of their side lobes. Broadening the main lobe allows for better beam overlap but does not allow the addition of additional beams. To achieve this apodization, it is required to modulate the amplitude of the output signal according to the position of the radiating element in the array. This can be achieved passively using attenuators or actively with variable amplification depending on the position of each element in the array mesh. However, this solution results in a reduction in the gain of the active antenna, for a given number of radiating elements, and is therefore not desirable.

[0012] In another approach described in the article “Reconfigurable Multi-Beam Pillbox Antenna for Millimeter Wave Automotive Radars” (M. Ettorre, R. Sauleau, Proc. ITST, pp. 87-90, 2009), sources are superimposed on two different levels, which however generates a significant coupling between the accesses.

[0013] There figure 3 illustrates the simplified operation of an E-plane combiner / divider, in which the sources are superimposed on two different levels (Port 1 and Port 2; Port 3 corresponds to the output port). Indeed, the odd-mode operation clearly highlights the poor isolation between the input ports and the poor adaptation of the excited input port (the E-field lines are not rectilinear).

[0014] US 2016 / 285165 A1 describes a compact Butler array consisting of a planar multilayer structure. US 2012 / 092224 A1 describes a multilayer antenna. US 7,724,197 B1 describes a parallel plate beamforming lens.

[0015] There is thus a need for improved quasi-optical beamformers capable of minimizing beam-crossing losses without significant increases in complexity and / or size. Résumé de l'invention

[0016] The invention is set forth in the attached set of claims. Description des figures

[0017] Other characteristics, details and advantages of the invention will emerge from reading the description given with reference to the appended drawings given by way of example. There figure 1 illustrates an antenna comprising a quasi-optical beamformer according to the state of the art. The figure 2 illustrates the radiation pattern for different depointing angles, with a state-of-the-art quasi-optical beamformer. The figure 3 illustrates several schematic representations of the operation of a state-of-the-art E-plane combiner. The figure 4 illustrates a top view (parallel to the XY plane) of the quasi-optical beamformer according to one embodiment of the invention. The figure 5 illustrates a perspective view of the quasi-optical beamformer, according to the section of the figure 4 . There figure 6 illustrates a perspective view of an embodiment of the port arrangement of the quasi-optical beamformer according to the invention, in which the ports are offset from each other. The figure 7 illustrates several schematic representations of the operation of a plane E combiner according to an embodiment of the invention. The figure 8 illustrates the radiation pattern for different depointing angles, with a quasi-optical beamformer according to one embodiment of the invention. The figure 9 illustrates a perspective view of an embodiment of the beam port arrangement of the quasi-optical beamformer according to the invention, comprising four stages of beam ports. The figure 10 illustrates a schematic representation, in the XZ plane, of an embodiment of the beam port arrangement of the quasi-optical beamformer according to the invention, comprising four stages of beam ports. The figure 11 illustrates a perspective view of one embodiment of the beam port arrangement, wherein the beam ports have different dimensions. figure 12 illustrates a perspective view of an edge of the quasi-optical beamformer according to an embodiment of the invention, comprising absorbers. The figure 13 illustrates a perspective view of an edge of the quasi-optical beamformer according to one embodiment of the invention, including dummy ports. The figures 14 , 15 And 16 illustrate different embodiments of implementing network ports and / or beam ports. The figure 17 illustrates a perspective view of the network ports of the quasi-optical beamformer according to one embodiment of the invention, in which the network ports are alternately connected to a load not connected to the antenna.

[0018] According to an embodiment of the invention illustrated by the figures 4 et 5 , the quasi-optical beamformer comprises an upper parallel plate waveguide 2 and a lower parallel plate waveguide 3, superimposed on each other. They thus share a common conductive plane 4, which constitutes the bottom wall of the upper parallel plate waveguide 2, and the top wall of the lower parallel plate waveguide 3. The upper and lower parallel plate waveguides extend in the XY plane, so they are superimposed along the Z direction.

[0019] The upper and lower parallel plate waveguides are not superimposed over the entire extent of the quasi-optical beamformer, but only over a part of it. Beyond a certain distance from the focal array of beam ports, the upper parallel plate waveguide 2 and the lower parallel plate waveguide 3 form, in the absence of a metal plane, a common parallel plate waveguide 5.

[0020] The quasi-optical beamformer also includes a set of upper beam ports 6 for feeding the upper parallel plate waveguide 2. The upper beam ports 6 are located in the plane of the upper parallel plate waveguide 2.

[0021] Similarly, the quasi-optical beamformer includes a set of lower beam ports 8 for feeding the lower parallel plate waveguide 3. The lower beam ports 8 are located in the plane of the lower parallel plate waveguide 3.

[0022] The quasi-optical beamformer also includes a set of network ports (7, 9), which can be arranged on a single level, in order to transmit the signals to the radiating elements.

[0023] The upper beam ports 6 and the lower beam ports 8 are located in the focal plane of the quasi-optical device 10. Each beam port comprises a source for generating a TEM (Transverse Electromagnetic) wave, a TE (Transverse Electric) wave, or both.

[0024] According to one embodiment of the invention, the sources are horns, in particular H-plane horns, which are particularly suitable for carrying out beam reconfiguration, each source of the beam port defining a spot access.

[0025] However, it should be noted that other well-known source forms can be used (monopole arrays, transitions between microstrip lines and parallel plate guides, transitions between tri-plate lines and parallel plate guides, transitions between coaxial guides and parallel plate guides, etc.). Horns can be easily designed and manufactured in PCB technology.

[0026] According to another embodiment, the quasi-optical beamformer comprises a single stage of beam ports, a set of upper network ports 7, and a set of lower network ports 9.

[0027] At the junction between the upper waveguide and the lower waveguide on the one hand, and the common waveguide on the other hand, a resistive film is arranged in the continuity of the conductive plane which separates the upper waveguide and the lower waveguide, as illustrated in figure 5 .

[0028] The resistive film is a layer that has a squared resistivity such that when current lines pass through the resistive film, some energy is dissipated, which reduces the coupling between the beam ports.

[0029] According to embodiments, the resistive film 11 may be closer to the beam ports than to the quasi-optical device, or conversely be closer to the quasi-optical device than to the beam ports. Similarly, the resistive film may be more or less wide (the width corresponds to the dimension along the longitudinal direction X).

[0030] Alternatively, the resistive film 11 may be adjacent to the beam ports, and / or adjacent to the network ports, i.e. in direct connection with the ports. In this case, the beamformer comprises only a single parallel plate waveguide, on a single stage.

[0031] It is possible to define the dimensions as well as the characteristics of the resistive film 11 by means of empirical measurements carried out during a simulation phase or during a calculation phase, so as to obtain the desired level of decoupling between the beam ports.

[0032] The dimension of the resistive film, in the propagation direction X, can advantageously be greater than or equal to λ g / 4, where λ g denotes the guided wavelength in the quasi-optical beamformer 1.

[0033] The resistive film may comprise, for example, a nickel-phosphorus alloy.

[0034] It is advantageous to arrange the resistive film 11 along the entire length of the metal plane 4, in the transverse direction Y, so as to dissipate the energy even for the beam ports most eccentric, relative to the main axis of the quasi-optical device.

[0035] The presence of the resistive film, in the continuity of the conductive plane (either directly in contact with the beam ports or the network ports, or at the junction between the superimposed guides and the common parallel plate waveguide), makes it possible to minimize losses linked to beam crossing.

[0036] Furthermore, the presence of the resistive film at the level of the superimposed beam ports (adjacent, or at the junction with a common parallel plate waveguide) makes it possible to free up space for the size of the sources, so that they perfectly illuminate the network ports, with an apodized law, also making it possible to reduce the secondary lobes. Larger sources also make it possible to limit the amplitude of the field on the edges of the quasi-optical beamformer, and to minimize parasitic reflections on them.

[0037] According to one embodiment of the invention, the beam ports (6, 8) and the network ports (7, 9) are superimposed on at least two levels (33, 34).

[0038] According to another embodiment, illustrated by the figure 6 , the upper beam ports 6 and the lower beam ports 8 can be offset relative to each other in the transverse direction Y, by a predefined distance. The offset is therefore carried out in the focal plane of the quasi-optical device 10.

[0039] The predefined distance is advantageously equal to the width of the beam port divided by the number of stages (33, 34) of beam ports, which makes it possible to obtain a compact network of beam ports.

[0040] Thus, as illustrated by the figure 6 , for a beamformer comprising two stages (33, 34), the predefined distance is equal to half the width of the beam port (d 2 / 2, d 2 corresponding to the width of a beam port) and the center of an upper beam port coincides with the junction between two lower beam ports, and vice versa.

[0041] There figure 7 illustrates, schematically, the operation of the quasi-optical beamformer according to the invention, at the junction between the upper parallel plate waveguide 2 and the lower parallel plate waveguide 3 on the one hand, and the common parallel plate waveguide 5 on the other hand.

[0042] The resistive film 11 makes it possible to isolate the upper 6 and lower 8 beam ports, and to obtain, at the output port 24, located in the common parallel plate waveguide 5, the lossless summation of the signals coming from the input beam ports when they are in phase and of the same amplitude (diagram a) of the figure 7 ).

[0043] Indeed, in the balanced (or even) mode the electric potential on either side of the resistive film 11 being identical, there is no current line created in the resistive part.

[0044] On the other hand, in the case of an imbalance between the input signals (odd mode, diagram b) of the figure 7 ), the resistive film 11 is subjected to current lines which lead to the absorption by dissipation of the imbalance between the input signals.

[0045] The resistive film 11 thus makes it possible to resolve the coupling problems that can be found in the state of the art.

[0046] There figure 8 illustrates the radiation pattern of an active multi-beam antenna comprising a quasi-optical beamformer according to the invention, in which the beam ports are superimposed on two levels. The active multi-beam antenna also comprises a radiating panel connected to the output of the beamformer. The abscissa represents the misalignment angle of the antenna.

[0047] The beam port number (1 to 22), visible on the right side of the figure which represents the quasi-optical beamformer, is found in the main lobe number in the left side of the figure. With the quasi-optical beamformer according to the invention, the crossover level is approximately 2 / 3 dB, which greatly minimizes the losses related to beam crossover, in comparison with the 9 dB observed when the beam ports are located on a single level.

[0048] The resistive film 11 thus makes it possible to adapt the upper and lower parallel plate waveguides to the common parallel plate guide, while ensuring low mutual coupling between the sources.

[0049] With such a level of overlap, the trainer according to the invention thus guarantees high-speed transmissions between satellites and fixed or fast-moving users (trains, planes, etc.).

[0050] The level of overlap can be further improved by increasing the number of floors, for example by arranging the beam ports on four floors.

[0051] Thus, according to an embodiment illustrated by the figure 9 , the quasi-optical beamformer comprises more than two stages, in this case four stages (33, 34, 35, 36). A resistive film (37, 38, 39) is arranged between each stage, adjacent to the beam ports. The beam ports of two superimposed stages can advantageously be offset by a predefined distance equal to the beam port width divided by the number of beam port stages. It can also be provided, in a configuration with four or more stages illustrated by the figure 10 , that the length of each conductive plane (41, 42, 43) along the direction X of propagation of a wave in the quasi-optical beamformer 1, is variable from one stage to another, so as, for example, to balance the coupling between the beam ports, in a progressive manner.

[0052] For example, the conductive plane 42 located at mid-height is the longest among all the conductive planes. Considering the stages located between the upper part 44 of the waveguide and the median conductive plane 42, the conductive plane located at mid-height 41 is assigned a length less than that of the median conductive plane 42, and so on (dichotomy division). The resistive films (111, 112, 113) are arranged at the ends of the conductive planes (41, 42, 43).

[0053] This embodiment ensures balanced coupling between the beam ports, and good distribution of the E field in even mode.

[0054] According to a particularly advantageous embodiment, the quasi-optical beamformer according to the invention is produced in the form of a multilayer PCB printed circuit. The permittivity ε r dielectric materials integrated into the beamformer in fact make it possible to reduce the guided wavelength inside the quasi-optical beamformer by a factor ε r , and to reduce the dimensions of the trainer by the same factor. The quasi-optical device 10 is integrated into a dielectric-loaded parallel plate guide, and the beam ports can be made using SIW (Substrate Integrated Waveguide) technology.

[0055] The manufacturing process of the quasi-optical beamformer thus includes a step of etching the resistive film, at the locations where the resistive film is planned. The manufacturing technique of a quasi-optical beamformer in PCB lends itself particularly well to the addition of a resistive film in the former.

[0056] Quasi-optical beamformers in the form of multilayer PCBs can cause more losses than beamformers in the form of metal guides. However, for active antennas, the amplifiers are integrated into the radiating panel (all amplifiers contribute to beamforming); they are therefore not integrated before the beamformer, which allows more tolerance for losses.

[0057] According to one embodiment of the invention, illustrated by the figure 11 , the dimensions of the beam ports are different from one floor to another. In this case, the number of beam ports is different from one floor to another. For example on the figure 11 , stage 37 comprises three beam ports 70, and stage 38 comprises four beam ports 71. The beam ports of stage 37 are wider (along the transverse direction Y) than the beam ports of stage 38. A portion of resistive film 11 extends at the junction between stage 37 and stage 38, at the output of the beam ports.

[0058] The embodiment illustrated by the figure 11 can be extended to more than two floors, for example four floors or more, with a conductor plane length that is fixed or variable from one floor to another.

[0059] The front of the cylindrical waves excited by the beam ports of the quasi-optical beamformer are oriented towards the barycenter of the array ports. The transmitted electric field is therefore maximum at the center of the array ports, and the electric field intensity may decrease for ports located at the periphery. However, there is a residual electric field at the edges of the quasi-optical beamformer.

[0060] In order to reduce the residual electric field at the edges, the quasi-optical beamformer, as illustrated by the figure 12 , comprises, on its lateral edges (25, 26), a first absorption device 12 in the upper stage 33, and a second absorption device 13 in the lower stage 34. The lateral edges (25, 26) are the edges located in the transmission line, between the beam ports and the quasi-optical device ( figure 4 ).

[0061] The absorption devices are configured to absorb the energy not transmitted between the beam ports (6, 8) and the network ports (7, 9), and thus to minimize the parasitic reflections on the edges of the quasi-optical beamformer.

[0062] . The first absorption device 12 and the second absorption device 13 may extend over the entire length of the corresponding lateral edge, namely entirely between the most eccentric beam ports and the quasi-optical device. Alternatively, the absorption devices may extend from the resistive film 11 to the quasi-optical device 10, along the longitudinal direction X.

[0063] The position of the first absorption device 12 and the second absorption device 13 is advantageously offset by a distance corresponding to λ g / 4 in the transverse direction Y, where λ g denotes the guided wavelength in the quasi-optical beamformer 1. The direction of the offset, i.e. which absorber is set back from the other, is not important. Furthermore, the resistive film 11 is arranged between the first absorption device 12 and the second absorption device 13. The resistive film 11 may extend beyond the absorption devices, in the transverse direction Y. The resistive film 11 may be arranged in the continuity of the metal plane and between the first absorption device 12 and the second absorption device 13, as illustrated in figure 12 .

[0064] The offset of the position of the first absorption device 12 and the second absorption device 13 by a distance corresponding to λg / 4 in the transverse direction Y generates a phase opposition between the parasitic reflections from the absorbers. The signal resulting from the combination in phase opposition is absorbed by the resistive film 11.

[0065] The reduction of parasitic reflections on the lateral edges (25, 26) makes it possible to limit the levels of signals interfering with the desired amplitude and phase laws on the network ports and thus attenuate the levels of the secondary lobes of the antenna.

[0066] The absorption devices may comprise an absorbent material, for example an epoxy foam loaded with magnetic particles.

[0067] According to a variant illustrated by the figure 13 , absorption devices may include dummy ports 33. Each dummy port may be in the form of a structure provided with a portion of resistive film 71, conductive side walls 72, and a conductive transverse link 70 which extends on either side of each side wall.

[0068] According to another variation, the absorption devices may comprise a plurality of dummy ports loaded with resistive loads.

[0069] There figure 14 illustrates an alternative arrangement of the network ports, in which the network ports 50 of a stage 33 are configured to be all coupled to an antenna, and the network ports 51 of an adjacent stage 34 are configured to be all coupled to a load 52 not connected to the antenna, which may be a resistive film. Coupling to a load 52 not connected to the antenna may be achieved by using horns connected to loads via transitions between rectangular guides and microstrip lines 53.

[0070] Another arrangement variant is illustrated by the figure 15 . The network ports on two levels use transitions between parallel plate guides and coaxial guides 54. The ports 56 of one of the two levels are connected to loads 55, which may comprise a resistive film. The ports 57 of the adjacent level are connected to the antenna.

[0071] Another arrangement variant is illustrated by the figure 16 . The network ports on two levels use transitions between parallel plate guides and microstrip lines 57. The ports 60 of one of the two levels are connected to loads 58 (for example, resistive films). The ports 59 of the adjacent level are connected to the antenna.

[0072] These different types of ports and transitions can also be used for beam ports.

[0073] This arrangement makes it possible to reduce parasitic reflections at high incidences and to use network port widths greater than 0.6λ g . Typically, network ports with widths less than 0.6λ g are used to limit these parasitic reflections.

[0074] Indeed, the incident waves are partially reflected on the network ports of each stage. This reflection increases with the size of the network ports and the incidence of the wave. The partial reflections of each stage are then in phase opposition when the network ports are shifted by half a period. They are then absorbed by the resistive film.

[0075] This partial reflection cancellation works for port widths up to 0.8λ g or even 0.9λ g , in order to reduce the angle of incidence of the quasi-optical beamformer waves θ QO required to feed the antenna.

[0076] Indeed, the angle of incidence θ QO is directly linked to the spacing d 2 between the network ports through the following equation, θ rad being the antenna misalignment angle, d 1 the spacing between the radiating elements of the antenna, ε r2 being the permittivity of the quasi-optical beamformer: θ QO = sin − 1 d 1 d 2 ε r 2 sin θ rad

[0077] The spacing d 1 between the radiating elements of the antenna is imposed by the constraint of placing the antenna array lobes outside the antenna coverage.

[0078] Typically for an active antenna of a satellite in geostationary orbit having to operate at θ rad = ±8.7°, the spacing between the radiating elements is of the order of 3.1λ where λ denotes the wavelength in vacuum.

[0079] Thus, for the case of an active antenna operating in a geostationary orbit, increase the periodicity of the network ports by 0.6λ g at 0.8λ g allows to relax the wave incidence constraint inside the quasi-optical beamformer, from 51.4° to 38.5°, which seems less critical.

[0080] This is possible by creating two overlapping rows of network ports spaced at a period of 0.8λ g , while implementing a half-period shift between the two superimposed rows. Only one of the two rows of ports is then connected to the radiating elements, and the ports of the other row are connected to loads (cf. figures 14 , 15 And 16 ), which helps avoid specular reflections.

[0081] According to another embodiment illustrated by the figure 17 , the upper and lower network ports are configured to be alternately coupled, along the transverse Y direction, to an antenna and to a load not connected to the antenna.

[0082] Thus, the upper network port set alternatively includes an upper network port 27 connected to the antenna (not visible on the figure 17 ), and a network port 28 connected to a load that is not connected to the antenna.

[0083] Similarly, the lower network port set alternately comprises a lower network port 29 connected to a load that is not connected to the antenna, and a network port 30 connected to the antenna.

[0084] Considering two overlapping network ports (e.g. ports 27 and 29, or ports 28 and 30), only one of the two ports is connected to the antenna, the other being connected to a load not connected to the antenna.

[0085] This operation explained for a receiving antenna also transposes to the case of a transmitting antenna. In this case, a wave incident on the network ports for an oblique incidence is partially reflected in the direction of the network lobe. The partial reflections are then converted into an odd mode, which vanishes in the resistive film.

[0086] The invention also relates to an active antenna comprising the aforementioned quasi-optical beamformer, and a radiating panel connected to the output of the beamformer.

Claims

1. A quasi-optical beam former (1) comprising a set of beam ports (6, 8), a set of network ports (7, 9), a quasi-optical device (10) and at least one parallel-plate waveguide (2, 3, 5) extending between the beam ports (6, 8) and the network ports (7, 9), the beam ports (6, 8) and / or the network ports (7, 9) being superposed in at least two stages (33, 34), each of the at least two stages (33, 34) being separated by a conductive plane (4) common to two adjacent stages (33, 34), characterised in that the quasi-optical beam former (1) comprises a resistive film (11) disposed in the continuity of the conductive plane (4).

2. The quasi-optical beam former (1) according to claim 1, wherein the at least one parallel-plate waveguide is a plurality of superposed parallel-plate waveguides (2, 3), each superposed parallel-plate waveguide (2, 3) being disposed facing the beam ports (6, 8) and / or facing the network ports (7, 9) of a same stage (33, 34), wherein the adjacent superposed parallel-plate waveguides (2, 3) in the two adjacent stages share the conductive plane (4) common to the two adjacent stages, the former (1) further comprising a common parallel-plate waveguide (5), disposed in the continuity of the superposed parallel-plate waveguides (2, 3), the resistive film (11) being disposed in the continuity of the conductive plane (4) at the junction between each superposed parallel-plate waveguide (2, 3) and the common parallel-plate waveguide (5).

3. The quasi-optical beam former (1) according to claim 1, wherein the resistive film (11) is adjacent to the beam ports (6, 8).

4. The quasi-optical beam former (1) according to claim 1, wherein the resistive film (11) is adjacent to the network ports (7, 9).

5. The quasi-optical beam former (1) according to one of the preceding claims, wherein, each beam port (6, 8) having an identical width (d2) between two consecutive beam ports (61, 62) of the same stage, the beam ports (61, 62) of two adjacent superposed stages (33, 34) are shifted by the width of the beam port divided by the number of stages (33, 34) of beam ports.

6. The quasi-optical beam former (1) according to one of the preceding claims, wherein the beam ports are superposed in at least four stages (33, 34, 35, 36), the length of each conductive plane (41, 42, 43) in the direction of propagation of a wave through the quasi-optical beam former (1) being variable from one stage to the next.

7. The quasi-optical beam former (1) according to one of the preceding claims, wherein the beam ports (70, 71) have different dimensions, from one stage to the next (37, 38).

8. The quasi-optical beam former (1) according to one of the preceding claims, wherein, each network port (7, 9) having an identical width between two consecutive network ports of the same stage, the network ports of two adjacent superposed stages are shifted by the width of the network port divided by the number of stages of network ports.

9. The quasi-optical beam former (1) according to one of the preceding claims, wherein the network ports (50) of a stage (33) are configured to all be coupled to one antenna, and the network ports (51) of a superposed adjacent stage (34) are configured to all be coupled to a load not connected to the antenna.

10. The quasi-optical beam former (1) according to one of the preceding claims, comprising, on each of the lateral edges (25, 26), a plurality of absorbing devices (12, 13) configured to absorb energy not transmitted between the beam ports (6, 8) and the network ports (7, 9), said absorbing devices (12, 13) being superposed in the at least two stages (33, 34), the position of the absorbing devices (12, 13) being shifted by a distance corresponding to λg / 4, where λg designates the wavelength guided in the quasi-optical beam former (1), the resistive film (11) being disposed between the absorbing devices (12, 13) of two superposed stages (33, 34).

11. The quasi-optical beam former (1) according to claim 10, wherein the absorbing devices comprise dummy ports or an absorber.

12. The quasi-optical beam former (1) according to one of the preceding claims, wherein the network ports and / or the beam ports comprise coaxial lines, coaxial guides, striplines or micro-strips.

13. The quasi-optical beam former (1) according to one of the preceding claims, produced as a multilayer printed circuit board (PCB), the parallel-plate waveguide being filled with a dielectric, the beam ports being produced in SIW technology.

14. An active antenna comprising a quasi-optical beam former (1) according to one of the preceding claims, and a plurality of radiating elements connected to the output of said beam former (1).

15. The active antenna according to claim 14, wherein the dimensions of the network ports are smaller than the dimensions of the radiating elements.