Background of the Invention and State of the Art In mobile telephony systems, apart from traffic channels on which speech and other types of data are transmitted between a base station and a mobile station, so called control channels transferring different types of control information are also used.

Some of these control channels, like the traffic channels, transfer point-to-point in- formation between the base station and the mobile stations. Other control channels are used by the base station for communication with all mobile stations within a sector cell at the same time. This requires an antenna at the base station having a sufficiently wide beam in the horizontal plane to cover the whole sector in question.

Such a sector covering beam usually has a limited beam width in the vertical di- mension and thus forms a horizontal disk, a so called flat beam.

The range requirement for channels for point-to-point information is the same as for channels for point-to-multipoint information. In present systems therefore one and the same sector antenna is used for both these functions. Point-to-point information, however, would not have to be transmitted from the base station in such a way that all mobile stations in the sector can receive it. It is enough that the mobile station for which the information is intended can. The base station, therefore, might concen- trate the transmit power, even sideways, to the desired directions by using antennas having radiation patterns with narrow beams. If the same antennas are used for re- ception as well, a corresponding increase in the receiver sensitivity in the desired directions is achieved. This concentration of the transmit power and the receiver sensitivity can be used to increase the range and/or lower the power demands on the

transmitters of both the base station and the mobile station. Since the channel fre- quency reuse spacing may be reduced with this method, the total capacity of the mobile telephony system may also be improved in this way.

One perceivable possibility of creating several simultaneous narrow beams is using a Butler matrix connected to an antenna array. A Butler matrix is a completely pas- sive and reciprocal circuit comprising an interconnection of a number of hybrid couplers and either fixed phase shifting elements or transmission cables of varying lengths. A Butler matrix for an antenna of N elements, N being an integer number, usually a power of two, has N input ports and N output ports and therefore enables the generation of N narrow beams. A signal on one of the input ports to the Butler matrix results in signals on the output ports of the matrix of substantially the same amplitude but different phases. Each input port corresponds to a certain combination of phases on the output ports. Each one of these combinations generates a narrow beam from the antenna array. Since the antenna and the Butler matrix are com- pletely reciprocal the system works as well for reception as for transmission.

Using an antenna fed from a Butler matrix, a set of narrow beams may be achieved, in which each individual radiation pattern has nulls for each angle at which another radiation pattern shows a maximum power (if the power is normalized using the antenna gain of the element pattern). Narrow beams meeting this criterion are said to be mutually orthogonal. Using a Butler matrix in combination with an antenna array to achieve a set of narrow beams is previously known per se.

It would be possible to use a separate sector antenna or alternatively one of the col- umns in an antenna array for the wide beam function. The lower antenna gain for the wide beam function would then have to be compensated with a higher amplify- ing power. The antenna gain here denotes the relationship between the maximum radiation of an antenna and the radiation of an ideal omnidirectional antenna with no

loss, with the same supplied power. For example, an antenna array with eighth col- umns has an antenna gain that is 9 dB higher than a single antenna column or a sector antenna. This implies that the power amplification of the amplifier must be 9 dB higher to compensate for the lower antenna gain.

UK patent specification GB 2 169 453 discloses a method of generating a number of narrow beams with different directions and one wide beam covering the same area as all the narrow beams together using an antenna array. Here an electromagnetic lens of a so called Rotman type with parallel plates is used. On one side of the lens there are a number of beam ports and on the opposite side there are a number of an- tenna ports. Each one of these antenna ports is coupled, through an amplifying module, to an antenna element in an antenna array. Each beam port corresponds to one of the narrow beams in the prior art. Further, the lens is equipped with a sepa- rate connection, the position of which on the lens is adjusted so that the geometrical distances to the antenna ports cause the supplied signal power to this connection to be divided over the antenna ports in such a way that a wide beam is generated from the antenna array.

The electromagnetic lens is a spacious and expensive component that is not avail- able on the market. Also, the wide beam, as in the previously described cases, ob- tains a lower antenna gain than the narrow beams, which requires expensive addi- tional, separate amplification for the wide beam not to give a shorter range than that of the narrow beams.

Summary of the Invention It is, as mentioned above, desirable to enable the implementation of an apparatus and a method for the simultaneous generation, with one antenna apparatus, of a number of narrow beams and a wide beam, substantially covering the same area as is covered by the individual narrow beams together, thus achieving a sufficient

range for the desired wide beam function. The range of the wide beam must be sub- stantially the same as that of the narrow beams. The narrow beams have a higher antenna gain compared to the wide beam function. Meeting these requirements have been a problem in the past.

The present invention solves this problem by utilizing an antenna array comprising a first number of sub-arrays each comprising at least one antenna element, and beam forming apparatus connected to the antenna array, such as a Butler matrix compris- ing a second number of antenna ports and a third number of beam ports, the activa- tion of each of at least a number of said beam ports separately corresponds to a ra- diation pattern characterized by a narrow main beam from the antenna array. By the simultaneous activation of at least a number of said beam ports by the same signal with suitable phase shifts a superimposition of the radiation patterns corresponding to the respective activated beam port is achieved in such a way that a wide beam is generated.

In the beam forming apparatus said antenna ports and beam ports are mutually con- nected in such a way that an individual activation of the beam ports, through an amplifying module for each port, causes a signal distribution on the antenna ports, specific for each beam port and corresponding to a specific radiation pattern with a narrow main beam from the antenna array. To the beam ports of the beam forming apparatus amplifying modules are connected. By distributing a wide beam signal, preferably with an even power distribution and supplying it to the beam ports through the amplifying modules the antenna array is caused to generate said wide beam. The wide beam signal is then transmitted from the antenna array over a rela- tively large angular interval. With suitable phase relationships on the wide beam signal at the beam ports the beam forming apparatus is thus brought to concentrate the signal power mainly to one of said antenna ports. Thereby the signal will mainly be transmitted by one of said sub-arrays, each of which comprises at least one an-

tenna element. The beam width of the wide beam will thus be determined mainly by the individual radiation pattern of the sub-arrays. By using all the amplifying mod- ules simultaneously when generating the wide beam the lower antenna gain of the wide beam will be compensated by a corresponding higher amplification, giving the wide beam the desired range.

The wide beam function is achieved by a suitable choice of phase relationships be- tween the beam signals. In a preferred embodiment of the invention practically all the power is concentrated to one of said antenna ports and thus also to one of the sub-arrays in the antenna array. The radiation pattern thus has a wide and smooth main beam.

An object of the present invention is to achieve an apparatus and a method for, using the same radio antenna apparatus, simultaneously being able to generate a number of narrow beams and a wide beam substantially covering the same area as is covered by the individual narrow beams together.

Another object of the invention is to achieve an apparatus and method for mobile telephony systems for enabling the communication between base stations and mo- bile stations over narrow beams.

An advantage of the present invention is that all the amplifying modules may be used simultaneously in the generation of the wide beam to obtain a sufficient range.

Another advantage of the present invention is that an apparatus for generating, si- multaneously and with only one radio antenna apparatus, a number of narrow beams and a wide beam is achieved, which meets high demands on cost and space.

A further advantage of the present invention is that it enables the utilization of nar- row beams in mobile telephony systems, through which reduced interference and improved use of frequencies may be achieved.

The invention will be described in more detail in the following by means of em- bodiments and with reference to the appended drawings.

Brief Description of the Drawings Figure 1 is a block diagram illustrating a preferred embodiment of the invention.

Figure 2 shows a view of radiation patterns obtained by the embodiment shown in Figure 1.

Figure 3 is a connection diagram showing a Butler matrix, according to prior art, for the embodiment shown in Figures 1 and 2.

Figure 4 shows a view of an embodiment of the invention used in a cellular mobile telephony system.

Figure 5 is a sketch-like block diagram illustrating the principles of an embodiment of the invention with a two-dimensional Butler matrix.

Figure 6 is a block diagram of a base station 71 in a cellular mobile telephony net- work according to an embodiment of the present invention.

Figure 7a is a signal diagram showing the radiation pattern of the embodiment shown in Figures 1, 2 and 3.

Figure 7b is a signal diagram illustrating the wide beam function for the embodi- ment shown in Figures 1, 2 and 3.

Detailed Description of Preferred Embodiments Figure 1 illustrates a radio antenna apparatus 10 comprising an antenna array 3 comprising eight antenna elements 3a,. . . ,3h, a Butler matrix 2 and eight amplifying modules la,. . .,1h. The Butler matrix 2 in turn comprises eight antenna ports A1,...,A8, each connected to an antenna element 3a,...,3h, and eight beam ports

2Ll,. . ,2L8 Each of said eight amplifying modules la,. . .,lh comprises a first con- nection L1,...,L8 and a second connection, said second connections being con- nected to said eight beam ports 2L1, ,2L8 Figure 2 illustrates the main radiation pattern of this radio antenna apparatus 10. The radio antenna apparatus is arranged to generate eight narrow, partially overlapping narrow beams 4a,. . .4h. Individual activation of the beam ports generates a signal distribution, specific to each beam port, on the antenna ports, corresponding to a narrow beam from the antenna array in a specific direction. Further the radio an- tenna apparatus is to be able to generate a wide beam 5, covering substantially the same area as the eight narrow beams 4a,...,4h together.

According to a preferred embodiment of the invention the narrow beams 4a,. . .4h will be mutually orthogonal. Hereby each individual narrow beam's radiation pat- tern has nulls for each angle in which another radiation pattern has a maximal power (if the power is normalized using the antenna gain of the element pattern).

The Butler matrix 2 is shown in more detail in Figure 3. Between the beam ports and . ,2L8 and the antenna ports A1,...,A8 the Butler matrix 2 comprises, as is known in the art, a first set of hybrid couplers ....... ,2 1 d, a second set of hybrid couplers 23a,. . .,23d and a third set of hybrid couplers 28a,.. .28d in such a way that each beam port 2L1, ..,2L8 is connected to each antenna port Al,.. .,A8. Supplied signal power on one of the beam ports will be distributed substantially evenly over the antenna ports. Further the Butler matrix comprises a number of fixed phase shifting elements 22a,. . .,22d, 24, 25, 26, 27. The bandwidth of the Butler matrix depends on the implementation of the hybrid couplers and the phase shifting ele- ments. There are examples of Butler matrices having a bandwidth of up to an oc- tave.

The definition of a Butler matrix dictates a defined relationship between the beam ports and the antenna ports of the matrix. A number of ways to implement a Butler matrix are, however, disclosed in the literature. The present invention is also not limited to Butler matrices. Other types of matrices, for example a so called Blass matrix or an electromagnetic lens of, for example, Luneberg or Rotman type may be used as beam forming apparatuses.

To generate a wide beam with the antenna arrays 3 one of the antenna columns of the antenna array may be used. The lower antenna gain for the wide beam function would then have to be compensated for with a higher amplifier gain. For example, an antenna array of eight columns has an antenna gain 9 dB higher than a single an- tenna column. This implies that the amplifier must have a 9 dB higher power am- plification to compensate for the lower antenna gain.

As shown in Figure 1, the amplifying modules la,...,lh in the present invention are arranged at the beam ports 2L1, . ,2L8 of the Butler matrix on the transmitter side of the Butler matrix 2 instead of the common location in radar applications, at the an- tenna ports. The amplification of these amplifying modules is dimensioned so that the range requirement is met with one amplifying module and the antenna gain for one narrow beam. This implies that each of the narrow beams meets the range re- quirement.

The desired wide beam, designated as 5 in Figure 2, is generated according to the present invention in that the wide beam signal distributed over the beam ports 2Ll,. ..,2L8 is combined at the antenna ports A1,...,A8 in such a way that they are added in phase in one of the antenna ports while they are added in the other antenna ports in such a phase relationship that substantially full cancellation occurs. In this way the signal will be concentrated to one of the antenna ports Al,.. .,A8. Since all

the amplifying modules are used together in this way, the total power will be the sum of the contributions of all the amplifiers.

The average power for each power amplifying module is dimensioned so that each individual narrow beam will give a certain Effective Isotropic Radiated Power (EIRP). EIRP by definition corresponds to the output power multiplied by the an- tenna gain normalized to an ideal isotropic transmitter. When generating the wide beam function the part of EIRP originating from the antenna gain will decrease by a factor of approximately M, M corresponding to the number of antenna columns (eight in this embodiment). On the other hand the part of EIRP originating from the power amplification will increase by the same factor M, so that EIRP will be the same for the narrow beam and the wide beam.

In this example it is assumed that the distances between any two adjacent antenna columns in the antenna array 3 are equal, that is, the antenna array is a so called Uniform Linear Array (ULA) having M=8 antenna columns 3a,...,3h. For a wave arriving straight on, an array response vector a(O) is obtained according to the fol- lowing: in which 0 denotes the angle between the narrow beam in question and the direction that is perpendicular to the antenna array and d is the distance between two adjacent antenna columns normalized to the wavelength. This response vector a(0) describes how the signals at the antenna ports are related to each other. The relationship be- tween beam port signals and antenna port signals for a Butler matrix is suitable de- scribed, in a way known per se, by a transfer matrix B according to:

b(O)=Ba(O), in which b(0) is a vector comprising M elements. Each element of this vector corre- sponds to a certain radiation function for each of the beam ports. The transfer matrix B has the dimension (M x M) and describes the relationships between the signals on the beam ports and antenna ports of the Butler matrix. H denotes a Hermitian conju- gation, that is both transposition of the transfer matrix and complex conjugation of the respective matrix element.

Each column B(k) of the matrix B corresponds to an amplitude normalized array re- sponse vector for a value of the angle 0, specific to each column. These angles are selected in such a way that all columns are mutually orthogonal, that is: BHB=E, in which E denotes the unit matrix. This gives: (BH)-l =B.

The combined radiation function gtot(#) at excitation of several antenna ports, is ob- tained by superimposing the respective radiation function of the antenna columns according to g,, (8) = (oTbb(O), in which ob is the excitation vector at the beam ports 2L1,..,2L8. This may also be written as

gtot(#)=(#bTBH)a(#), in which the excitation of the antenna columns is obtained according to #cT = #bTBH, #b being the excitation vector at the beam ports 2L1,..,2L8. If the whole signal power is concentrated to a single antenna port the combined radiation function gtot(#) of the antenna array will be determined by the characteristic of a single antenna column, thus giving a wide beam. The excitation vector ob at the antenna ports is therefore set to be a vector Uk, an arbitrary vector element of the vector Uk being constituted by a constant C and all other vector elements being zero. This gives: (0bT =UTk(BH)-l = UkTB.

If, for example, the antenna port denoted as A2 in Figure 1 is to be excited, the fol- lowing function is obtained for the excitation vector C3b: It follows that the excitation vector ( b at the beam ports should be one of the rows of the transfer matrix B, in this example row 2, multiplied by a constant to concen- trate all the signal power to one of the antenna columns. Since all the matrix ele- ments ideally have the same value for a Butler matrix, this means that the beam ports of the Butler matrix should be excited by the same signal strength to obtain a

smooth, wide beam. The mutual phase of the beam port signals should coincide with an arbitrary row in the transfer matrix B.

According to an alternative embodiment of the invention, the phase of the wide beam signal is changed instantaneously at regular points in time at the beam ports of the Butler matrix, in such a way that the signal power from the wide beam signal is moved from one antenna column to another in the antenna array. By this procedure the power losses, and thus also the heating caused by power loss, are shared, reduc- ing the demand and increasing the lifetime.

In this example a Butler matrix is used as a beam forming apparatus, causing the narrow beams to be orthogonal. This fact has been used when deducting the excita- tion vector 0)b above, when it was shown, among other things, that the signal ampli- tudes in the beam ports 2LI,...,2L8 should ideally be equal. Orthogonality is, how- ever, no absolute prerequisite for the invention. If a beam forming apparatus that does not give absolute orthogonality is used, the elements of the excitation vector ob will, however, require different values for an even, wide beam to be obtained from the antenna array 3. The power amplifying modules la,...,lh must therefore supply different output powers, which impairs the link budget of the radio system. Accord- ing to a preferred embodiment the beam forming apparatus therefore provides or- thogonal or substantially orthogonal beams.

As the antenna array 3 and the Butler matrix are entirely reciprocal elements the same antenna can also be used for reception. The receiving function is suitably en- abled by means of a set of duplex filters between the amplifying modules la,...,lh and the Butler matrix 2.

In the embodiment shown here the wide beam signal is divided on the baseband side. It is, however, possible to modulate this signal separately, divide the modu-

lated wide beam signal and, after a suitable phase shift, feed it to said first connec- tions L1,. . .,L8 of the eight amplifying modules la,. .. lh.

A field of application for the radio antenna apparatus 10 is shown in Figure 4. In cellular mobile telephony systems so called sector cells are often used. In this case three base stations are placed in the same geographical location, usually referred to as a site, and have their respective antennas directed so that each antenna serves a sector cell of 120 degrees. In the figure, six such base station sites BS1,...,BS6 are shown. At the site BS4 a first base station serves a first cell C1, a second base sta- tion serves a second cell C2 and a third base station serves a third cell C3.

According to prior art the antennas at the base stations are characterized by wide beams covering an entire sector cell. Three wide beams B 1, B2, B3 covering the first cell C1, the second cell C2 and the third cell C3, respectively, are shown in the figure. With these wide beams the respective base stations can communicate with the mobile stations that are found within the cells. Such a mobile station MS is shown in the figure. A large part of the information that is exchanged between the base stations and the mobile stations consists of point-to-point information. It would, however, not be necessary to transmit such point-to-point information in such a way that all mobile stations within the sector can receive it. It is sufficient that the mobile station for which the information is intended can receive the signal.

The base stations in this embodiment of the invention use narrow beams for the point-to-point information. In this way, the output power may be concentrated to the desired directions. In the figure one such narrow beam P1 is shown. With this nar- row beam the mobile station MS communicates with the base station of the cell C2 in which the mobile station is located.

The higher antenna gain caused by the narrow beam in this way improves the link budget in both directions, that is, to and from the base station. This may be utilized

to increase the range relative to the output power of the base station and the mobile stations. The total capacity of the mobile telephony system may also be improved with this technology compared to the prior art, since the frequency reuse spacing may be reduced.

Some information transmitted by the base stations should, however, be received by all the mobile stations found in the cells concerned. The base stations according to the present invention are therefore able to generate wide beams. These should have substantially the same range as the narrow beams. Since each base station comprises a radio antenna apparatus, denoted as 10 in Figure 1, each base station can generate a number of narrow beams, which together cover the cell in question. At the same time the base stations can generate a wide beam which substantially covers the whole cell.

Figure 6 is a simplified overview of a transceiver, in this case a base station 71 in a cellular mobile telephony network, said transceiver comprising a radio antenna ap- paratus according to an embodiment of the present invention. The base station 71 is an example of a communication device comprising such a radio antenna apparatus.

Other types of communication devices may use such a radio antenna system in the same way.

The base station 71 comprises a baseband processing unit 4 connected to an in- put/output (I/O) unit 6. The base station 71 further comprises a radio antenna appa- ratus 10 like the one described in connection with Figure 1. The radio antenna appa- ratus 10 comprises an antenna array 3 comprising eight antenna elements, a beam forming apparatus in the form of a Butler matrix 2 and an amplifying unit 1 com- prising eight amplifying modules. Between the amplifying unit 1 and the Butler matrix 2 a duplex filter unit 9 is arranged, comprising a first, a second and a third set of connections. The amplifying unit 1 is connected to the first set of connections and

the Butler matrix is connected to the second set of connections. To the third set of connections a second amplifying unit 8 is connected. To this second amplifying unit 8 comprising eight amplifiers, a demodulator unit 7 is connected, which in turn is connected to the baseband processing unit 4. The baseband processing unit 4 is also connected to the input terminal of a modulator unit 5. On the output terminal of the modulator unit 5 the amplifying unit 1 is connected.

The duplex filter unit 9 is arranged, in a way known in the art, to separate the re- ceiver part of the base station, comprising said second amplifying unit 8 and de- modulator unit 7, from the transmitter part of the base station comprising the first amplifying unit 1 and the modulator unit 5.

Each amplifying module in the amplifying unit 1, the output of which is connected through the duplex filter unit 9 to a single beam port of the Butler matrix 2, is con- nected to a single modulator in the modulator unit 5. With this arrangement the sig- nal intended to be transmitted in a specific narrow beam is modulated separately. In a corresponding way the signal from each signal beam port in the Butler matrix 2 is demodulated separately in the demodulator unit 7. The signal demodulated in this way therefore originates from a single narrow beam.

When transmitting data to all the mobile stations in the base station's cell the ampli- tude of the signal is evenly distributed over all inputs of the modulator unit. Thus, all the amplifying modules in the amplifying unit 1 will be used in the amplification of this signal. When suitable phase relationships of the signals are used, the Butler matrix 2 will generate such a signal distribution over the antenna ports of the Butler matrix 2 that a wide beam will be generated from the antenna array 3.

The radio antenna apparatus described above is particularly well suited for mobile telephony systems using Single Carrier Power Amplifier (SCPA) technology (that

is, carrier specific amplifiers used in the base stations) when several different carri- ers are used at the same time. This requires that the signal to be transmitted is am- plified before different carrier waves are mixed. This requirement is met according to the present invention by the amplification being made on the beam port side of the beam forming apparatus, and thus before the combination of the carriers. Fur- ther, a radio antenna apparatus according to the present invention is particularly well suited for Spatial Division Multiple Access (SDMA) in which several active radio connections are used simultaneously on the same carrier but within different beams.

In the embodiments of the invention described above a one-dimensional Butler ma- trix is used. The term one-dimensional here implies that the control takes place in one dimension even if each antenna column in the antenna array in a preferred em- bodiment of the invention comprises several antenna elements. The invention is, however, not limited to control only in one dimension. In Figure 5 a principle sketch of a two-dimensional Butler matrix 50 is shown, by means of which the beams from an antenna array may be controlled in two dimensions. The two-dimensional Butler matrix 50 comprises a first set of one-dimensional Butler matrices ....... ,5 1 f. The two-dimensional Butler matrix 50 further comprises a second set of one- dimensional Butler matrixes 52a,...,52h cascade coupled with said first set of one- dimensional Butler matrices ....... ,5 1 f.

Each Butler matrix 51a,...,51f in said first set of Butler matrices comprises eight beam ports and eight antenna ports. In a corresponding way, each Butler matrix 52a,...,52h in said second set of Butler matrices comprises six beam ports and six antenna ports. Each antenna port of the Butler matrices 52a,. . .,52h is connected to an antenna element in a two-dimensional antenna array 53. This antenna array 53 in this example comprises 6 x 8 = 48 antenna elements.

Each of the eight antenna ports of the Butler matrix 5 la, which are hidden in the Figure, is connected to one of the Butler matrices 52a,. . .,52h in said second set of one-dimensional Butler matrices. In the same way, each one of the Butler matrices 5 lib, . . .,Slfis connected to each Butler matrix 52a,. . .,52h in said second set of But- <BR> <BR> <BR> ler matrices. In this way, each antenna port of the matrices 5 1 ....... ,5 1 f is connected to one of the beam ports of the matrices 52a,. . .,52h.

With the first set of Butler matrices control takes place in a first dimension. With the second set of Butler matrices control takes place in a second dimension. In this way <BR> <BR> <BR> <BR> the activation of each one of the beam ports of the matrices 5 1 ....... ,5 1 fin said first set of Butler matrices corresponds to a radiation pattern from the antenna array.

A wide beam is generated according to this embodiment by the even distribution of the amplitude of a wide beam signal to the two-dimensional Butler matrix 50. This wide beam signal is power amplified by means of a set of amplifying modules not shown in the figure. With suitable phase relationships of the wide beam signal dis- tributed over the amplifying modules, the two-dimensional Butler matrix 50 is caused to concentrate the supplied signal power to substantially one single antenna port of an arbitrary matrix of the one-dimensional Butler matrices 52a,. . .,52h. In this way, the wide beam signal will mainly be transmitted by one of said antenna elements in the antenna array 53. The beamwidth of the wide beam obtained in this way will then mainly be determined by the individual radiation pattern of this an- tenna element.

The phase relationships of the wide beam signal distributed over the amplifying modules, are determined by the two-dimensional Butler matrix 50. It can be shown that 48 different phase relationships fulfil the criterion that theoretically all power is to be concentrated to an antenna port of one of the one-dimensional Butler matrices

52a,. . . ,52h. Each of these 48 phase relationships corresponds to a concentration of the signal power to one of the 48 antenna elements in the antenna array.

According to an alternative embodiment of the invention the phase relationship of the wide beam signal is instantaneously changed at regular points in time at the beam ports of the two-dimensional Butler matrix, in such a way that the signal power from the wide beam signal is moved from one antenna element to another in the antenna array. In this way the power losses, and the heating associated with power losses, are distributed over the antenna elements, reducing the demand and increasing the lifetime.

Figure 7a is a signal diagram showing a radiation pattern for the embodiment pre- sented above in connection with Figures 1, 2 and 3. In the signal diagram S denotes signal strength, measured in decibel, and 0 denotes an angle relative to the direction perpendicular to the antenna array. In the signal diagram eight radiation functions are illustrated, each characterized by a narrow beam 61,...,68 and a number of side lobes with a low amplitude compared to the narrow beam. The excitation of one of <BR> <BR> <BR> <BR> the beam ports of the Butler matrix, denoted 2LI ,. . .,2L8, in Figure 1, corresponds to one narrow beam 61,...,68 with associated sidelobes from the antenna array 3. Since the Butler matrix generates orthogonal radiation patterns, there are, as indicated in Figure 7a, angles in which all eight radiation functions except one substantially has the value zero.

Figure 7b is a signal diagram illustrating the wide beam function of the embodiment presented in connection with Figures 1, 2 and 3. When all eight beam ports, denoted 2Ll,...,2L8 in Figure 1 are excited with an even amplitude distribution and such phase relationships as discussed in connection with Figure 1, a wide beam 70 is ob- tained which substantially covers the same angular area as the narrow beams 61,...,68 in Figure 7a, taken together.