GB2367188A

GB2367188A - Shaped antenna beam

Info

Publication number: GB2367188A
Application number: GB0023483A
Authority: GB
Inventors: Leonard Ogier
Original assignee: OGIER ELECTRONICS Ltd
Current assignee: OGIER ELECTRONICS Ltd
Priority date: 2000-09-25
Filing date: 2000-09-25
Publication date: 2002-03-27
Also published as: AU2001286094A1; GB0023483D0; WO2002025776A1

Abstract

An antenna for a base station of a mobile phone network is mounted at the top of a tall tower. In order to obtain a constant level of received power as a mobile phone moves around, the power distribution across the beam emitted by the base station is varied with elevation so as to cancel out the inverse-square law effects due to moving away from the tower. This is achieved by using a phased array which is designed as follows. A first estimate of the array size and a number of elements is made based on the predetermined variation in radiated power with elevation and a desired azimuth beamwidth at a predetermined operating wavelength, selecting a maximum difference between desired and actual array gain, making a first estimate of the voltage and phase of the signal to be fed to each element of the array, calculating an elevation pattern for the first estimate for at least one azimuth angle, optimising the phase and power distribution between the elements by comparing the calculated elevation pattern with the desired predetermined pattern and if the difference is greater than the selected maximum difference, adjust the phase and/or voltage of the feed to at least one of the elements to reduce the difference and repeat the calculating and optimising steps until the difference is less than the selected maximum difference, calculating the loss of gain at maximum range compared to an equivalent sized array having elements fed in-phase with equal power, optimising the array size by repeating the voltage and phase estimating step, the elevation calculating step and the phase and power optimising step for different array sizes, an selecting the size having the least reduced gain.

Description

SHAPED ANTENNA BEAM

This invention relates to an antenna and to a method of designing an antenna particularly, though not exclusively for use in point to multipoint ground-based radio systems.

Point to Multipoint radio systems are being used increasingly. Examples include mobile telephone networks, rural telephones and microwave video distribution. These systems comprise one or more hub stations to which a number of subscribers communicate. A new generation of point to multipoint systems is also being introduced to provide high speed internet, business data and closed user group communications.

All these systems require simultaneous or near simultaneous communications to and from subscribers at short as well as long ranges. Typically the range from the hub to the subscriber can be from 50 metres up to 10 km or more. This means that the signal strength will be 46 dB (40,000 times) greater at close range that at long range.

The signal strength will also vary from one hub transmitter to another, and from one subscriber transmitter to another. Similarly the performance of the antennas will vary, as will the sensitivity of the receivers. Overall, these effects can easily add a further 10 dB (10: 1) variation.

Rain and other atmospheric effects can introduce variations in signal strength. This factor is becoming increasingly important because the low frequencies are becoming occupied and newer systems have to use higher frequencies which are more susceptible to rain fade. The amount of fade that has to be accommodated is a function of the availability requirements, e. g. a service required for 99.999% of the time has to tolerate more fade than one providing 99.9% availability. Typically the fade at 5 km can vary from 4 dB at 10 GHz to 24 dB at 40 GHz.

When all these factors are taken together the received power variation or dynamic range can be 60 to 80 dB.

The conventional way of dealing with wide dynamic ranges is to design receivers to accommodate the large signal variation, sometimes combined with automatic gain control. The disadvantage of this is that it increases the cost and complexity of the receiver. When the total dynamic range is of the order of 60 dB or more, the implementation can be become difficult, especially at higher microwave frequencies.

Against this the subscriber equipment should be as simple and as inexpensive as possible. Thus the need for wide dynamic ranges and AGC is not attractive.

Another method of reducing the dynamic range is to shape the transmit beam at the hub to compensate for the variation of power due to range. As differences in range account for the largest variation in signal strength, if it can be eliminated, it will simplify the receiver design.

Beam shaping has been used previously in some applications, in particular in airborne radar applications, to provide a constant return signal from the ground at all ranges. In this case a cosecant squared beam shape is used.

To date, however, only rudimentary beam shaping has been employed in high frequency ground based, point to multipoint links. The extent of the shaping has been to direct the beam downwards to optimise the coverage area, and to"fill-in"the nulls nearest to the main beam of the antenna. Against this, the directivity requirements of airborne equipment are easier to achieve because the design is less onerous.

Figure 1 compares the radiation pattern required for an airborne radar with that required for a point to multipoint hub. The aircraft plot shown relates to an aircraft which is at 10,000 metres and requires a 100 km range. It can be seen that the beamwidth is greater than 2 degrees and the attenuation at 40 degrees off boresight is only 17 dB. Against this, a point to multipoint hub antenna installed at 30 metres and required to provide coverage to 5 km, has a very much sharper beam. The beamwidth is a fraction of a degree and the attenuation at 40 degrees is greater than 40 dB.

The simpler requirements of the airborne system allow approximations to be used in the design. In one approach, Fourier transformation of the required beam shape is employed to calculate the voltages and phases in an array of radiating elements.

The limitations of this approach are that the Fourier analysis assumes an infinite number of radiating elements whereas in practice all practical antennas have a finite size.

In the second approach the radiation pattern is calculated at a number of discrete angles assuming a continuous aperture rather than an array of discrete points.

There are two approximations in this approach. Firstly, the pattern between the angles is not calculated. Secondly, a continuous aperture is assumed rather than a discrete number of elements.

Because the requirements of a ground based, point to multipoint, system are appreciably more onerous, greater precision is required in the design process than can be provided using the above techniques. A further complication is that if that if the design requires a relatively narrow azimuth beam, then an M x N array of elements is required. The above procedures do not apply to these designs.

According to the invention there is provided a method of optimising the design of a multi-element antenna array to achieve a predetermined variation in radiated power with elevation comprising making a first estimate of the array size and number of elements based on the predetermined variation in radiated power with elevation and a desired azimuth beamwidth at a predetermined operating wavelength, selecting a maximum difference between desired and actual array gain, making a first estimate of the voltage and phase of the signal to be fed to each element of the array, calculating an elevation pattern for the first estimate for at least one azimuth angle, optimising the phase and power distribution between the elements by comparing the calculated elevation pattern with the desired predetermined pattern and if the difference is greater than the selected maximum difference, adjust the phase and/or voltage of the feed to at least one of the elements to reduce the difference and repeat the calculating and optimising steps until the difference is less than the selected maximum difference, calculating the loss of gain at maximum range compared to an equivalent-sized array having elements fed in-phase with equal power, optimising the array size by repeating the voltage and phase estimating step,

the elevation calculating step, and the phase and power optimising step for different array sizes, and selecting the size having the least reduced gain.

According to a second aspect, the invention provides a ground-based antenna array comprising a plurality of discrete radiating elements such as patch elements and a feed network arranged to feed each patch element with a predetermined feed voltage and at a predetermined phase, the feed network being arranged such that the elevation beam shape of the antenna has substantially the same nominal radiated power at all ranges.

The number of elements, the array size and the feed network are preferably arranged by making a first estimate of the array size and number of elements based on the predetermined variation in radiated power with elevation and a desired azimuth beamwidth at a predetermined operating wavelength, selecting a maximum difference between desired and actual array gain, making a first estimate of the voltage and phase of the signal to be fed to each element of the array, calculating an elevation pattern for the first estimate for at least one azimuth angle, optimising the phase and power distribution between the elements by comparing the calculated elevation pattern with the desired predetermined pattern and if the difference is greater than the selected maximum difference, adjust the phase and/or voltage of the feed to at least one of the elements to reduce the difference and repeat the calculating and optimising steps until the difference is less than the selected maximum difference, calculating the loss of gain at maximum range compared to an equivalent-sized array having elements fed in-phase with equal power, optimising the array size by repeating the voltage and phase estimating step, the elevation calculating step, and the phase and power optimising step for different array sizes, and selecting the size having the least reduced gain.

The method and antenna may be used for a ground based antenna to be used at the hub of a point to multipoint system which has an elevation beam shape to provide the same nominal power at all ranges. A particular feature of such an antenna is that it typically uses an array of discrete patch elements to achieve the beam shape.

The design principles involved are different to those employed on other systems. It may use a computerised analysis to calculate the radiation pattern directly and to optimise the voltages and phases within the array in an iterative routine to minimise the error between the actual and the wanted pattern whilst simultaneously maximising the gain of the antenna.

In other aspects, the invention also provides a computer program embodying the method aspect of the invention and a computer readable data carrier carrying the program.

Embodiments of the invention will now be described by way of example with reference to the drawings in which : Figure 2 is an elevation of a typical antenna array using circular polarised transmission; Figure 3 is a chart showing the voltage level and phase of signals to be fed to each element of a typical 32 element antenna array; and Figure 4 shows a practical embodiment of the antenna of Figure 3 using unequal power dividers and different feeder line lengths.

A typical array using circular polarised transmission is illustrated in Figure 2. The array has patch elements 2 (not all referenced) fed by transmission lines 4 (not all referenced) coupled using 2-way dividers 6 (not all referenced). In this case, the azimuth beamwidth is typically 100 degrees whilst the vertical beam is matched to provide constant power at all ranges from an 80 metre tower.

Design The stages of the design process are: 1 Calculate the approximate size of the antenna 2 Select the maximum error between the required and actual shapes 3 Compute the voltages and phases required at each element

4 Calculate the loss of gain from an standard array 5 Repeat the above steps for different size antennas 6 Translate into a practical design Step 1. Antenna Size Knowing the height of the hub in a point to multipoint system relative to subscribers, it is possible to calculate the required radiation pattern in the elevation plane. Some approximations are required because the surrounding terrain will not be flat.

For the sake of this illustration we assume that the azimuth coverage is broad and can be met by a vertical array of N patch elements. Narrower azimuth beams may require additional horizontal patches H in an N x H (for example 2 or 4 in an N x 2 or N x 4 array) array The illustration of Figure 1 shows the required pattern if everything were perfect. In practice if there were errors in the actual design, these will introduce ripples In the response which could reduce the signal at short range below that received at the maximum possible range. Also if there are valleys below the average ground level, even lower levels will be received.

Both these effects force the design into increasing the power levels from the ideal pattern to provide some contingency. 20% is recommended. As an example, if the theoretical design requires 30 dB loss at a specific elevation angle, the actual design could be for 24 dB loss.

From this pattern, the beamwidth can be calculated and then the first estimation of the optimum size obtained. The relevant equations are: Beamwidth = 2 x (angle of max signal-angle where signal is 3 dB lower) Antenna size = 70 x Wavelength/Beamwidth in degrees The optimum separation between patches or other similar elements is typically three quarters of a wavelength. Small separations provide little benefit in the overall gain

whereas larger separations often result in high sidelobes. On this basis, the initial estimate for the number of patches is : 1.33 x Antenna size/Wavelength Because feed networks generally use simple power splitters from one into two ports, this results in arrays that use multiples of 2 elements, i. e. 2, 4 8, 16,32 etc. This is not invariably the case if more complex splitters can be employed, but for this example we assume that only simple 2 way devices are to be used. This means that the number of patches has to be rounded up to the nearest value which is multiple of two.

The above procedure establishes the starting point for the vertical array. If the azimuth beamwidth is narrow, then several elements may also be required in the horizontal direction.

The skilled person will appreciate that the choice of additional horizontally disposed vertical arrays may be made in the way taught in"Antenna Engineering Handbook" (Third Edition). Edited by Richard C Johnson and published by McGraw-Hill, Inc.

(the disclosure of which is incorporated herein by reference).

Step 2. Select the maximum error The maximum gain of the antenna in the coverage area reduces as the design is optimised to minimise the difference between the wanted and the actual beam pattern. This is because the best match is frequently obtained by off-setting the wanted pattern from the peak.

Consequently there is a trade-off between maximum antenna gain (at the maximum range of the system) and a good match between the actual and wanted patterns. This means it is not advisable to over-specify the degree of optimisation, especially since local reflections and other effects will in any case degrade the theoretical performance.

Typically, 2 to 3 dB difference between the actual antenna pattern and the wanted pattern (as modified to take account of ripple and varying topology) is recommended. Using this, the loss of gain from an array of similar size is less than 2 dB.

Step 3. Compute the voyages and phases at each element Having established the number of elements and their spacing it is now possible to calculate the radiation pattern of the antenna using the teaching of'Antenna Engineering Handbook' (Third Edition). Edited by Richard C Johnson and published by McGraw-Hill, Inc., for example For reasonable accuracy, it is recommended that the pattern should be calculated for at least 20 points in the elevation plane. Initially it is assumed that each element is fed with the same voltage and that all patches are fed in phase.

It should be noted that if the patches or other radiating elements are non-isotropic over the sector in which the pattern is calculated, then the amplitude and phase characteristics of the elements has to be included in the overall array calculations.

However in the case of a single vertical array of simple elements, it is necessary only to calculate the elevation pattern at one azimuth angle. The performance of the array at other azimuths can then be deduced by superimposing the azimuth patterns of the elements on a common elevation shape. Alternatively if the antenna comprises 2 or more horizontal elements within the vertical array, it will be necessary to calculate the elevation pattern at several azimuth angles. Three or more azimuth cuts may be required.

The pattern calculated above, which is that of a uniformly fed array, can then be compared to the wanted pattern and the point at which maximum difference between the two can be determined. Initially this will be between two of the nulls because at this point the signal will have its maximum value but will be in anti-phase to the wanted signal.

Knowing the difference and the elevation angle at which it occurs, it is possible to calculate the voltage and phase to be added to each of the elements to reduce the difference to zero. For example if the difference or error from 32 isotropic elements is 32 mV at 30 degrees, then 1 mV will have to be added to each element with a relative phase relationship than provides vector addition at 30 degrees offset.

In practice it is found that the system provides more rapid optimisation if half the error voltages are added to the elements rather than the full error voltage. When this has been done, the radiation pattern can then be re-calculated and the new angle at which there is a maximum difference from the wanted pattern can be established.

From this, the further voltage and phase increments for each patch can be calculated.

The above optimisation process continues until the maximum error becomes equal to, or less than, the maximum tolerable error.

Step 4. Calculate the loss of gain The loss of gain at the maximum range can be determined by calculating the signal power from the array at the elevation angle which corresponds to the maximum range and comparing it to the power from an array of the same size assuming that the same input power were equally distributed in-phase to all the patches.

Step 5. Repeat for different size antennas The procedure described in steps 3 and 4 is then repeated for different arrays lengths, keeping the number of elements constant. Typically the length should be varied from 70% to 140% of the original length in 10% steps.

Following this, the loss of gain can be plotted against array length and the optimum, which provides the minimum loss, can be selected. It should be noted that the feeder loss will increase as the array length increases. An allowance should be included for this. Step 6. Translate into a practical design The voltage level and phase of the signals to be fed to each of the elements is illustrated in Figure 3 for a typical antenna. This can be realised using conventional microwave design techniques by using a number of unequal power dividers together with different line lengths. The configuration is illustrated in Figure 4.

Claims

1. A method of optimising the design of a multi-element antenna array to achieve a predetermined variation in radiated power with elevation comprising : (a) making a first estimate of the array size and number of elements based on the predetermined variation in radiated power with elevation and a desired azimuth beamwidth at a predetermined operating wavelength, (b) selecting a maximum difference between desired and actual array gain, (c) making a first estimate of the voltage and phase of the signal to be fed to each element of the array, (d) calculating an elevation pattern for the first estimate for at least one azimuth angle, (e) optimising the phase and power distribution between the elements by comparing the calculated elevation pattern with the desired predetermined pattern and if the difference is greater than the difference selected in step (b), adjust the phase and/or voltage of the feed to at least one of the elements to reduce the difference and repeat steps (d) and (e), (f) calculating the loss of gain at maximum range compared to an equivalent-sized array having elements fed in-phase with equal power,

(g) optimising the array size by repeating steps (c) to (f) for different array sizes, and selecting the size having the least reduced gain as calculated in step (f).

2. A method according to claim 1, wherein the step of making a first estimate of the array size includes calculating the array height substantially according to the formula 70 x (operating wavelength/beamwidth in degrees) where beamwidth is defined as 2 x (angle of maximum signal angle at which signal is 3dB less than maximum).

3. A method according to claim 2, wherein the number of elements in a set of vertically aligned elements distributed generally evenly over the length of the array is calculated substantially according to the formula 4/3 x array size/operating wavelength.

4. A method according to claim 3, including introducing a plurality of substantially identical sets of the vertically aligned elements laid out side-by side, the number of sets depending on the desired beamwidth.

5. A method according to claim 4, wherein the number of sets is divisible by two.

6 A method according to any preceding claim, wherein the said maximum difference is in the range 1 to 4dB and preferably in the range 2 to 3dB.

7. A method according to any preceding claim, wherein the first estimate of voltage and phase has all voltages the same and all phases the same.

8. A method according to any preceding claim, wherein the elevation pattern is calculated for a plurality of sample points having the same azimuth angle and distributed in elevation angle.

9. A method according to any preceding claim, wherein the elevation pattern is calculated for a plurality of sample points having a plurality of predetermined azimuth angles and distributed In elevation angle.

9. A method according to claim 8 or claim 9, wherein the pattern is calculated for twenty or more sample points at each azimuth angle.

10. A method according to any preceding claim wherein the comparison between the calculated elevation pattern and the desired predetermined pattern is made by finding the elevation angle of the point at which the maximum difference occurs between the calculated and desired patterns.

I

11. A method according to claim 10, wherein the power fed to the array is adjusted at each element by an amount up to the maximum difference divided by the number of elements and wherein the phases of the feed to each of the elements are adjusted to provide vector addition at the elevation angle of the point of maximum difference.

12. A method according to any preceding claim, wherein the array sizes chosen in step (g) are chosen from a range of between 70% and 140% of the first estimate of array size.

13. A method according to any preceding claim, wherein the comparison of different array sizes includes an allowance for feeder loss for different length arrays.

14. A ground-based antenna array comprising a plurality of discrete radiating elements such as patch elements and a feed network arranged to feed each element with a predetermined feed voltage and at a predetermined phase, the feed network being arranged such that the elevation beam shape of the antenna has substantially the same nominal radiated power at all ranges.

15. An antenna according to claim 1, wherein the number of elements, the array size and the feed network are arranged according to the following design process (a) making a first estimate of the array size and number of elements based on the desired elevation beam shape and a desired azimuth beamwidth at a predetermined operating wavelength, (b) select a maximum difference between desired and actual array gain, (c) make a first estimate of the voltage and phase of the signal to be fed to each element of the array, (d) calculate an elevation pattern for the first estimate for at least one azimuth angle, (e) optimise the phase and power distribution between the elements by comparing the calculated elevation pattern with the desired predetermined pattern and if the difference is greater than the

difference selected in step (b), adjust the phase and/or voltage of the feed to at least one of the elements to reduce the difference and repeat steps (d) and (e), (f) calculate the loss of gain at maximum range compared to an equivalent-sized array having elements fed in-phase with equal power, (g) optimise the array size by repeating steps (c) to (f) for different array sizes, and selecting the size having the least reduced gain as calculated In step (f).

17. A computer program embodying the steps of any preceding method claim.

18 A computer readable data carrier carrying the program of claim 17.

19. A method of designing an antenna as described herein with reference to the drawings.

20. An antenna constructed and arranged as described herein with reference to the drawings.