Description FREQUENCY AND POLARISATION SELECTIVE
MULTIBEAM ANTENNA.
Background to the Invention.
[1] The invention relates to antennas and in particular to the problem of selecting microwave signals of different frequencies, bandwidth and polarisation. It relates also particularly to multi-beam antennas.
[2] In many communications or radar applications electrically large antennas are used, i.e. they are large compared to the wavelength of the radio or microwave signal. Such antennas are directive, i.e. they have a narrow beam and exhibit gain. Antennas using reflectors which are illuminated by a primary feed - typically a waveguide horn - are very common and well established. A very common type of reflector antenna is the parabolic dish which is often used for satellite applications such as broadcast television. A dish antenna however generally only works well with a single primary feed, since the parabolic dish exhibits a single focus, and hence a single beam can be used with a single source of signals e.g. a geostationary satellite. It is possible to add further primary feeds in proximity to the focus of a parabolic dish, producing additional beams over a narrow range of scan angles e.g. about 20 degrees, but efficiency will reduce with increasing scan angle.
[3] Where multiple antenna beams are required over a wider range of solid angles, e.g.
180 degrees (a hemisphere of space), a reflector antenna is generally not suitable. A solution is offered by the Luneburg lens antenna, which is a spherical dielectric lens with varied refractive index. A variant uses a dielectric hemisphere lens in conjunction with a reflective (metal) plate. Such an antenna can produce multiple beams by using multiple feeds placed around the outside of the lens, or a mechanically steered beam could be used. Further variants may use a constant refractive index lens (where the focussing efficiency tends to be low), or a number of discrete dielectric layers (where the efficiency can be high).
[4] Where different frequencies or polarisations are required, it is usually the feed design which controls the frequency and polarisation response. Considering for example satellite communications, it is common to use a feed which operates over the frequency range 10 - 13 GHz (broadcast television typically exists within this range). Where some other service is offered on another frequency, for example 20 GHz or 30 GHz, it becomes quite problematic to make the feed work over all these frequencies. In general, two frequency bands can be achieved with relative ease, but it is more difficult to work over three different bands. A technique to overcome this problem is en- countered with the frequency selective surface (FSS), which allows a beam to be split in space according to frequency. Then, a number of discrete feeds may be used in the region close to the focal point of an antenna: the technique is well established for reflector antennas. However, if a multiple frequency (e.g. 12 GHz, 20 GHz, & 30 GHz) hemisphere lens antenna is required, the use of a FSS is problematic because the feed is generally placed close to the edge of the lens and there is a lack of space for placing the FSS components. A solution to a similar problem has been put forward whereby a spherical or Luneburg lens is divided by an FSS and the two feeds placed on opposite sides of the FSS - this requires for a full spherical lens structure. The concepts apply equally to discrimination of polarisation as well as frequency.
[5] The present invention overcomes the problem of separating out frequencies (or polarisations) using a hemisphere lens geometry. It arises, in part, from the observation that it is convenient to arrange for the different signals to emerge from the lens structure at different locations, but that it is not particularly critical how those locations are related. It is sufficient that, for example, two foci are separated in space by a distance sufficient to allow two primary feeds to be placed side-by-side without their colliding. This then allows for existing primary feed designs (these typically being single band or dual band) to be used, and obviates the need to deploy a tri-band feed. The invention relates to a modification of the hemisphere lens ground plane region so as to spatially separate different signals. Brief Description of Drawings.
[6] Figure 1 illustrates a side view of a first embodiment of a single beam antenna comprising a lens, feed and selective surface.
[7] Figure 2 illustrates a side view of a second embodiment of a single beam antenna comprising a lens, plurality of feeds, selective surface, dielectric wedge and ground plane.
[8] Figure 3 illustrates a side view of a third embodiment of a multiple beam antenna comprising a lens, plurality of feeds, selective surface, dielectric wedge and ground plane.
[9] Figure 4 illustrates a side view of a fourth embodiment of a single beam antenna comprising a segmented lens, plurality of feeds, two frequency selective surfaces of opposite type, two ground planes.
[10] Figure 5 illustrates a side view of a fifth embodiment where additional dielectric wedges, selective surfaces and feeds are added to the structure described under the second embodiment.
[11] Figure 6 illustrates a short dielectric cylinder or spherical segment, various dielectric wedges which may be formed from such a cylindrical or spherical segment, and a stratified dielectric sphere whose radial dielectric distribution closely resembles that of the wedges.
[12] Figure 7 illustrates a top view of an antenna array where the array element is one of the hemispherical lens antennas with feeds and selective surfaces, and where the elements are connected to a power distribution circuit, and shows the geometry for beam steering in one axis for example azimuth.
[13] Figure 8 illustrates a side section of an antenna array illustrated in Figure 7 and shows a geometry for beam steering in one axis, for example, in elevation.
[14] Figure 9 illustrates a modified dielectric and reflective plane region which is stepped, or zoned, so as to reduce the maximum thickness of the dielectric layer.
[15] Figure 10 illustrates the stepped dielectric region of Figure 9 in more detail by illustrating a region where material is removed.
[16] Figure 11 illustrates sketches of the stepped ground planes and dielectric regions of
Figure 9. Detailed Description of Embodiments.
[17] Referring to Figure 1, an antenna 1,2,5,20 comprises an electromagnetic dielectric lens 1 which is either hemispherical or approximately hemispherical or a segment of a hemisphere, at least one first antenna feed 5 and at least one frequency and/or polarisation selective surface element 2 located on the plane face of the hemisphere lens.
[18] The frequency and/or polarisation selective element 2 may be a frequency selective surface, or may be a polariser, or may be a surface or stratified group of surfaces which combines
[19] these functions. Such a surface is typically realised as a periodic surface comprising a metalised pattern as documented by 'Frequency Selective Surfaces', by B.A.Munk, published by Wiley & Sons Inc. 2000.
[20] The hemispherical lens may be a Luneburg lens, a constant-index lens, or a stepped- index lens, and when used in conjunction with a reflective surface 2 acts to yield a focus at the location of the feed 5.
[21] Where the selective surface 2 is a frequency selective surface, it serves to reflect a signal of a chosen frequency bandwidth which is illustrated as following the path 9-10-5, where 9 is the direction of maximum gain of the antenna or boresight, 5 is the antenna feed, and 10 is the centre of the hemisphere lens plane outer surface. Clearly, rays exist also in all other parts of the hemisphere lens, and the above serves to illustrate the relationship between the angles of the antenna boresight, feed, and selective surface.
[22] Where the selective surface 2 is a frequency selective surface, those signals outside of the frequency range at which the selective surface acts as a reflector, i.e. frequencies at which the surface is approximately transparent, are not reflected by the surface and pass into region 20 i.e. they follow the path 9-10-21 and those signals may be further manipulated in region 20.
[23] Referring to Figure 1, in a first embodiment of the invention the region 20 comprises an absorptive material, in which case the frequency and polarisation response of the antenna, and hence that of the radar cross section of the antenna, is dominated by the frequency and polarisation response of the selective surface 2. In certain applications, e.g. a stealth antenna, this property can be advantageous, since the radar cross section can be minimised by careful design of the selective surface, for example a very narrow band frequency response. This overcomes a problem encountered with electrically large reflector antennas which have a large radar cross section which can be undesirable. In the case of reflector antennas, an equivalent technique would entail the fabrication of a frequency and/or polarisation responsive reflector, but this can lead to structural problems because the reflector is in general a curved, three dimensional surface whereas in the present invention this difficulty is ameliorated considerably because the selective surface is a plane, readily fabricated, and a supportive structure is offered by the plane face of the dielectric lens.
[24] Referring to Figure 2, in a second embodiment of the invention the region 20 comprises a further dielectric region 4 and a reflective surface 3. The dielectric region 4 is a wedge shaped segment from a cylinder, or a spherical segment, whose radial dielectric profile matches that of the hemisphere lens 2 i.e. it may have a constant index or stepped index. A second antenna feed 6 is then added at an angular distance b where it is clear from principles of geometrical optics that the angle b equals twice the angle a, where the angle a is the angle between the selective surface 2 and the reflective surface 3. This invention then comprises a multi-frequency and multi- polarisation antenna, and overcomes the problem of manufacturing wide -bandwidth antenna feeds. In a variant of the embodiment, the hemisphere lens 1 is approximately hemispherical, i.e. this allows that when lens 1 and wedge 4 are electromagnetically coupled they form a combined lens which is hemispherical.
[25] The invention differs from 'multi-beam antenna' invented by Ebling, James P and
Rebeiz, Gabriel (publication numbers WO2004010534, WO0137374) in that said invention comprised a pair of hemisphere lens antennas in conjunction with a selective surface, while the present invention comprises a single hemisphere lens and at least one selective surface and combinations of other components which may include: absorptive region, a reflective plane and dielectric wedges. Thus the present invention, for a given electrical aperture size, occupies approximately half the physical volume of 'multi-beam antenna' of Ebling and Rebeiz and also is more easily mounted on a plane surface due to the essentially hemispherical geometry.
[26] By way of example, and so as to describe the useful properties of the invention, we may consider an antenna for satellite communications where the antenna is required to work at several frequency bands, for example in Ku band typically 10-14 GHz, and at K or Ka band, typically where a link at close 20 GHz is required e.g. a downlink from a satellite and a link at close to 30 GHz is required e.g. an uplink to a satellite.
[27] Here, a problem has traditionally been encountered whereby it is difficult to construct an antenna feed which can discriminate the three bands. Typically, two bands can be discriminated, e.g. using a wide-band horn feed and waveguide orthomode transducers, for example to discriminate 20 GHz and 30 GHz. To discriminate between Ku band e.g. 10-14 GHz and a higher band e.g. 20 GHz or 30 GHz, a dielectric rod feed may successfully be used. This however still does not solve the problem of discriminating a third frequency band. An established solution to this problem is found in reflector antennas where a frequency selective surface is added between the reflector and a plurality of feeds, but such an antenna typically has very limited scanning ability. A similar approach to use a frequency selective surface with a hemisphere lens, i.e. where the surface is added in the spatial region between the hemisphere lens and the plurality of feeds, is complicated by the lack of available space in this region. The invention overcomes this problem. A further advantage of the invention is that it has very wide scanning ability, i.e.: a plurality of feeds may be placed at various points around the outer edge of the hemisphere lens so as to produce a plurality of beams in the manner of a Luneburg antenna.
[28] Continuing the theme of the above example, a single antenna terminal may be required to communicate with a satellite at the three frequency bands described (Ku band and 20 GHz and 30 GHz) and may also be required to communicate with a second satellite spaced at an angular distance away from the first satellite and using either the same three frequency bands or a subset of these three frequency bands, or indeed some other bands. The particular advantages of the invention become apparent when it is noticed that the antenna could utilise commercially available feeds, for example Ku band Low Noise Block feeds, and Ka band transceiver feeds where these combine transmit and receive functionality, for example respectively at 30 GHz and 20 GHz. It may also be noticed that the exact angle b between the feeds and hence the signals which are spatially discriminated by frequency, is not particularly critical, and needs only to be sufficiently large to place the two feeds without their colliding with one another, hence the dielectric layer 4 needs be made only as thick as necessary to achieve this spatial separation of the feeds and this dimension may typically be of the order of 50 millimetres since a typical antenna primary feed has such an aperture diameter.
[29] Referring to Figure 3, in a third embodiment of the invention, an additional feed 13 or plurality of feeds 13 and 14 are added at an angular distance c away from the first feed 5 or plurality of feeds 5 and 6. In such a case, the feeds now act in clusters, i.e. where feeds at 5 and 6 serve to generate a combined antenna beam at spatial direction 9 and the feeds at 13 and 14 serve to generate a combined antenna beam at spatial direction 15. The relative angular separation between feed cluster 5 and 6 and feed cluster 13 and 14 is the same as the angular separation c between antenna beams 9 and 15. These locations may be chosen arbitrarily and typically in practice to generate beams at pre-determined directions and by way of example the beams point at two or more satellites in geostationary orbits. The scanning ability of the antenna, i.e. its ability to produce focussed beams over a wide range of angles, is a feature of the hemisphere lens. The selective surface allows different frequencies or polarisations to be discriminated from within a beam which points in a given direction. This selectivity is a consequence of the geometry of the selective surface 2 combined with the dielectric wedge 4 and reflective plane 12.
[30] Referring to Figure 4, in a fourth embodiment of the invention, the antenna comprises spherical lens segments 20, 21 and 26. Segment 20 is substantially hemispherical but is not a complete hemisphere. Segments 21 and 26 are also segments from a hemisphere. Surfaces 22 and 27 are reflective surfaces. The segments are so arranged that when segment 20 is electromagnetically coupled to segment 21 and elec- tromagnetically isolated from segment 26, a complete hemispherical electromagnetic lens is formed. In this case, antenna feed 5 yields a main radiation lobe at direction 9. The segments are also so arranged that when segment 20 is electromagnetically isolated from segment 21 and electromagnetically coupled to segment 26, a complete hemispherical electromagnetic lens is also formed and in this case antenna feed 6 yields a main radiation lobe at direction 9. The electromagnetic coupling or isolation between the segments is achieved using selective surfaces 23 and 24, and where these are of opposite type i.e.: a signal exhibiting a certain frequency bandwidth or a certain electromagnetic polarisation to which surface 23 is transparent, should be reflected by surface 24. This signal is also reflected by the surface 22. Then, surfaces 22 and 24 form a continuous reflective plane surface which delineates the planar surface of the hemisphere comprising segments 20 and 21. The signal which experiences focussing by the combined hemisphere lens 20 and 21 is associated with a main radiation lobe at direction 9 and the antenna feed 5.
[31] Conversely, a signal exhibiting a certain frequency bandwidth or a certain electromagnetic polarisation to which surface 24 is transparent, should be reflected by surface 23. This signal is also reflected by the surface 27. Thus, surfaces 23 and 27 form a continuous reflective plane surface which delineates the planar surface of the hemisphere comprising segments 20 and 26. The signal which experiences focussing by the combined hemisphere lens 20 and 26 is associated with a main radiation lobe at direction 9 and the antenna feed 6. [32] It is apparent that the angle b between the feed 5 and the feed 6 is twice the angle d between the plane of surface 22 and surface 24 and the plane of surface 23 and surface 27.
[33] Surfaces 23 and 24 are contiguous along the line 25 which is normal to the plane of
Figure 4 and passes through the origin of symmetry of the lens structure. Similarly, reflective surfaces 22 and 27 are contiguous along line 25.
[34] To explain the functionality of the fourth embodiment an example is described. In one example, surface 23 is transparent to a particular frequency range, for example 20-30 GHz, and surface 24 is reflective to this frequency range. In such a case, feed 5 yields a radiation lobe at direction 9 for this frequency range. However, should surface 23 be reflective for another frequency range, for example 10-14 GHz, and surface 24 transparent to this frequency range, feed 6 yields a radiation lobe at direction 9 for this frequency range. This example functionality is equivalent to that described in embodiments 2 to 4 i.e. discrimination of certain frequencies (or polarisations, or both frequency and polarisation) and which allows the use of existing (or future) antenna feeds. An advantage of the fifth embodiment compared to embodiments 2 to 4 is that the signals which are discriminated exhibit focussing by an electromagnetic lens which retains a hemispherical shape for both discriminated signals, whereas in embodiments 2 to 4 it is apparent that at least one of the discriminated signals experiences focussing by an approximately hemispherical lens.
[35] In a fifth embodiment of the invention, the antenna comprises a hemispherical lens or a variant as described under the first embodiment or a variant as described under the fourth embodiment, and a plurality of dielectric wedges and selective surfaces as described under the second embodiment or dielectric wedges and selective surfaces of opposite type as described under the fourth embodiment, and a plurality of feeds, where each feed is associated with one of the selective surfaces and where each feed operates over a certain range of frequencies and polarisations which are reflected by the selective surface associated with that feed. Feeds may be used singularly or in clusters in any number or combination so as to generate multiple antenna beams. While any number of dielectric wedges and selective surfaces may be used, an example is shown in Figure 5, where dielectric lens 1 and selective surface 2 are associated with feed 5 which forms a main radiation lobe at direction 9 as described under the first embodiment. Dielectric wedge 4 and selective surface 30 are associated with feed 6. Dielectric wedge 32 and selective surface 33 are associated with feed 31. Further dielectric wedges, selective surfaces, and feeds may be added. It is clear that the angles between any two feeds e.g. b and g are twice the value of the angles between the selective surfaces associated with the feeds e.g. respectively, angles a and/.
[36] By way of further illustrating the shape and form of the dielectric wedges described above (4 in Figure 2 and in Figure 3, 21 and 26 in Figure 4, 4 and 32 in Figure 5) Figure 6 shows sketches of the various types, where these may take the form of segments of a short cylinder or spherical segment 40.
[37] The short cylinder or spherical segment 40 may have uniform dielectric constant or radially stepped dielectric constant or continuously radially varied dielectric constant.
[38] Where segment 40 is a segment of a sphere, said sphere may have uniform dielectric constant, or radially stepped dielectric constant, or continuously radially varied dielectric constant in the manner of a Luneburg lens. A case where the sphere has stepped dielectric constant is shown by way of example as concentric dielectric layers 51,52,53,54. The segment 40 would then be that illustrated as a cross section through layers 51,52,53,54, i.e. region 61 in Figure 6. The cross-section of region 61 is delineated by points 62,63,64,65. It is apparent that for segment 40, which is characterised also by dimensions height h and diameter D, where h is less than D, the segment 40 resembles, to a close approximation, a short cylinder. In such a case (h is less than D ) segment 40 may be more easily manufactured from cylindrical i.e. disc- like sections rather than spherical layers 51,52,53,54. This observation is described in some detail to show that while the ideal geometry of the hemisphere lens antenna retains spherical surfaces which are continuous at dielectric boundaries, or boundaries between dielectric materials and air, the dielectric wedges may be more easily fabricated where they take the form of cylindrical or disk- like components.
[39] The dielectric wedges are illustrated in Figure 6 as sketches showing perspective: wedge 41 corresponds to wedge 4 in Figure 2 and in Figure 3, and to wedges 4 and 32 in Figure 5. Also in Figure 6 wedge pairs 42 and 43 correspond to wedge pairs 21 and 26 in Figure 4. In Figure 4 line 25, which is normal to the plane of Figure 4, corresponds to line 44-45 in Figure 6.
[40] In a sixth embodiment of the invention, feeds or clusters of feeds in any of the combinations described in embodiments 1 to 5 are mechanically actuated and/or electrically switched so as to change the direction of the antenna boresight. This gives rise to a steerable antenna beam, which may combine properties of both electronic and mechanical steering, such as may be useful to maintain a communications or radar signal between locations which are moving with respect to each other. To illustrate the usefulness of the invention, such examples might be a fixed ground terminal which communicates with a satellite in a non-geostationary orbit, or a moving vehicle which communicates with a satellite or a terrestrial base station or another vehicle, and in particular where a plurality of antenna beams and radio frequencies are required to be discriminated.
[41] In a seventh embodiment of the invention, and referring to Figure 7 and to Figure 8, a plurality of the antennas described in embodiments 1 to 5 may be combined so as to produce an antenna whose aperture area, and therefore aperture gain, is greater than that of each individual antenna. Such a technique, or device, is commonly called an array antenna and is a well established technique within the microwave and antenna industry. In this case, the individual antenna is called an antenna element, or element, and the input/output from each element is combined using a power combining network typically to combine all the elements' powers at a single input/output port.
[42] To illustrate the array principle Figure 7 shows four antenna elements each comprising 70,71,72 of the type described in embodiments 1 to 5 mounted above selective surface 73. In this example, two feeds 71 and 72, which discriminate between two states of electromagnetic polarisation or frequency according to the second embodiment, are associated with each dielectric lens 70. Each antenna feed 71 has a port, or connector, where at a microwave transmission line may be connected. Such transmission lines 602 may be combined, for example as shown, in a parallel combiner network via power dividers 600 to yield a single input/output port 601. The second group of antenna feeds 72 may similarly be combined.
[43] The usefulness of the seventh embodiment becomes apparent when the following functionalities and advantages are elucidated. A high gain, single beam antenna can be formed from an array of individual antennas. The higher gain thus offered allows for higher performance communications links than would be the case for single antenna elements. Such an antenna would be particularly useful for a mechanically steered antenna, such as might be used on a vehicle. In such a case, where the antenna is for example placed on the upper surface or roof of a vehicle, Figure 7 then represents a top view of the installation and the azimuth steering angular coordinate is shown where 75 is the pivot point which would typically be placed in the centre of the array antenna. The fabrication costs of the antenna are likely to be advantageous compared to the fabrication of a single lens antenna of equivalent aperture area. The height of the antenna can be less than that of a single lens or reflector antenna. These advantages have also been identified by M. Rayner in, 'Use of Luneburg Lens for Low Profile Applications', by Datron/Transco Inc. Microwave Product Digest, December 1999, where a similar array of four lens elements was reported for a single beam antenna. The present invention however puts forward advances to the 'Datron' product reported by Rayner, whereby the selective surfaces of embodiments 1 to 5 can considerably expand the functionality of the array antenna. Conveniently, the selective surface may be fabricated as a single continuous surface 73, and any lower reflective surfaces may similarly be fabricated each as a single continuous surface. The side view of the seventh embodiment, where it is used as a steerable antenna, is illustrated in Figure 8. While any of the embodiments 1 to 5 may be used as the lens antenna element, Figure 8 shows, by way of example, a geometry for the element substantially equivalent to the second embodiment, i.e. lens 70 is placed above selective surface 73 above dielectric wedge 75 above reflective plane 74 where an antenna beam at spatial direction 76 is discriminated into electromagnetic signals of two frequencies or polarisations (or combinations thereof) at feeds 71 and 72. Now, where said components 70,73,75,74 are together rotated in elevation, and feed elements 71 and 72 remain fixed with respect to a chassis structure 85, elevation scanning of the antenna beam will take place. More specifically point 80 is the centre of the hemisphere lens planar face and by way of example this is illustrated in Figure 8 also as being the pivotal axis for the elevation steering mechanism. Another pivotal axis could be chosen. In the example of Figure 8, line 84-80-82 represents the vertical centre line for the non-steered condition and the antenna beam falls at direction 76. Pivot 80 is a line running perpendicular to the plane of the diagram. When a mechanical elevation steering angle s is applied between vertical line 80-82 (non-steered condition) and line 80-83 (steered condition), that is, all of components 70,73,75,74 are rotated clockwise about pivot 80, the antenna beam moves to elevation direction 81 i.e. it scans through elevation angle t. It is apparent from geometrical optics that antenna scan angle t is double the value of the angle of mechanical movement s.
[44] As stated above variants of the seventh embodiment are possible for example where the electromagnetic lens components of the fourth embodiment are used as the antenna element: here it would be advantageous to place the central axis 25 in Figure 4 coincident with the elevation scanning axis 80 of Figure 8.
[45] Further advantages specific to the seventh embodiment are now highlighted.
[46] On elevation scanning, antenna feeds 71 and 72 are not moved with respect to a fixed chassis structure 85. This is in contrast to the 'Datron' antenna where the lens antennas are fixed with respect to a base plate and the antenna feeds, with the associated power combiner, are moved with respect to the lenses, and on scanning to high elevation angles the feeds hence partially block the aperture, leading to loss. Thus, an advantage of the present invention is that due to the offset feed geometry (feeds 71 and 72 in Figure 8 are placed an angular distance away from the vertical position 80) the antenna can scan to the zenith without the feeds incurring aperture blockage.
[47] A further advantage of the present invention is that high performance, low loss rigid waveguide can be used for the power combiner network. This is shown schematically as waveguides 90 and 91, which respectively combine power from/to arrays of feeds 90 to transceiver unit 92, and array of feeds 71 to transceiver unit 93. An important feature is that waveguide components and feeds are all fixed rigidly to a chassis structure 85. Chassis structure 85 is rotated in azimuth to perform azimuth beam steering. A flexible cable, rotary joint, or a wireless technique may then be used to connect radio transceiver units 92,93 to another part of the vehicle on which the antenna is mounted.
[48] The antenna described in the seventh embodiment can discriminate different frequency bands and/or polarisations in a manner substantially equivalent to embodiments 1 to 5. Beam steering is achieved by mechanically steering in elevation the lens/selective surface sub-chassis with respect to a main chassis 85, and by rotating chassis 85 in azimuth. The ability to use rigid waveguide power combiners is advantageous where high microwave frequencies are used (e.g. greater than 20 GHz) and where large arrays are required (e.g. greater than 2 elements).
[49] In a variant of the seventh embodiment, additional feeds may be placed in proximity of the lens edge, each at a pre-determined focal position, so as to produce a multi-beam steerable antenna. In this case, the antenna feeds are fixed with respect to each other and produce beams which similarly have a fixed angular relationship with respect to each other. Such a multi-beam steerable antenna is particularly useful where a moving vehicle requires to maintain communications links with multiple geostationary satellites, since said satellites have an approximately constant angular relationship to each other when viewed from a ground station or moving ground station.
[50] In an eighth embodiment of the invention and referring to Figure 9, the selective functionality described in the second embodiment is achieved using a modified dielectric wedge region 102 and ground plane region 101. Here, the ground plane region 101 is constructed as a series of steps where each maintains the same angle b with selective plane 2. The purpose of this embodiment is to reduce the maximum thickness of dielectric region 102 in a manner similar to a 'zoned' or Fresnel dielectric lens. By way of example, three such 'zones' are illustrated. The selectivity arising from discrete antenna beams associated with feeds 5 and 6 is achieved in an equivalent manner as in the second embodiment, i.e. the angle b between the selective surface 2 and the majority of the area of the ground plane region 101 gives rise to the angular separation of the two antenna beams.
[51] The geometry is described in more detail in Figure 10 which shows by way of example the three dielectric wedge and ground plane zones of Figure 9. The zoned dielectric region 102 is a section taken from the original dielectric wedge 4 of Figure 2, whose dielectric distribution is explained in Figure 6 (the wedge 4 is a section from the short dielectric cyliner 40 in Figure 6.) In Figure 10, dielectric region 102 can be envisaged as resulting from the removal of volume 402 (drawn hatched in Figure 10) from wedge 4. Viewed in cross-section, the segment cross-section 401 is repeated for each of the three zones, each of which is associated with a lower ground plane (reflective plane) 403. Each ground plane zone 403 maintains the same angle with selective surface 2 of Figure 2. The key feature of the eighth embodiment is that maximum dielectric zone thickness t is less than the maximum thickness s for the non- zoned equivalent: this feature reduces antenna mass and minimises the deviation from hemispherical geommetry that is introduced by the dielectric wedge 4 in Figure 2 i.e. the difference in path length in the wedge region experienced across different parts of the wedge region is reduced in the eighth embodiment compared to the second embodiment.
[52] While the relative dimensions in Figure 10 and the number of zones are arbitrary, it is worth listing some 'rules of thumb' which should be self-evident to those skilled in microwave antenna design:
[53] The length of the sloped ground plane zone 403 should preferrably be at least several wavelengths at the frequency of operation.
[54] and:
[55] The change in dielectric thickness at each step, which is the same as the 'height' (as viewed in Figure 11) 403 at each ground plane step, should preferrably be less than one wavelength at the frequency of operation.
[56] To further illustrate the concept of the zoned ground plane region, Figure 11 shows further sketches of the components, where surface 500 is a selective surface which is placed conformally to the plane face of a dielectric hemisphere lens, and a non-zoned reflective surface 501 is placed at an angle with respect to the selective surface 500, as described above for the second embodiment. The geommetry for the zoned case is illustrated where by way of example three reflective zones 502 are placed each at the same angle with respect to selective surface 500. The zoned dielectric regions 503 are inserted in the region between the selective surface 500 and the zoned reflective planes as arrowed in the sketch. The upper surface 505 is a continuous surface which is then conformal to the selective surface 500. The zoned dielectric region may be fabricated as a single, combined component 504, or as discrete components 503. The lower surfaces of dielectric components 503 are evidently discontinous (illustrated here by way of example as three plane surfaces oriented with respect to selective upper surface 505 by a common angle) and these discontinuous surfaces are conformal to the reflective surfaces 502. The sketches in Figure 11 are illustrative and intended to aid understanding.