The Development of Wide Angle Base Station Antennas from Arrays
Use of flat panel array base station antennas to produce wide angle beamwidths up to 120°
Phased array antennas are traditionally used to produce narrow beams from large apertures. In this mode of operation, every radiating element in the array is broadly co-phased.
The radiation pattern of the antenna is then the product of both the individual radiating element characteristic pattern and the array factor. The latter tends to be a sharply peaked function, which causes the resultant antenna radiation pattern to be sharply peaked in the desired direction. Figure 1 shows a typical radiation pattern for a classical phased array antenna. Base station antennas, sometimes referred to as sector antennas, often have a requirement for the azimuth beamwidth of the antenna to be broader than the characteristic beamwidth of a single radiating element. When this is the case, there are two methods to achieve the required beamwidth. The first is to use a different radiating element with a broader characteristic beamwidth. The second is to use a phased array in which the elements are driven with largely differing phases. These phases can produce a beamwidth broader than the individual element, but also offer the possibility to shape the antenna beam. This allows a rapid roll-off in radiated signal strength outside the desired sector angle. This characteristic is very advantageous in point-to-multipoint systems in which it is often a requirement to re-use the same frequency for different sector antennas pointing in different directions on the same site. The use of phased arrays offer other benefits, too. For example, the resulting antenna will be flatter and more discreet than an antenna using alternative technologies. Also, by allowing the use of a single printed circuit board (PCB) on the final antenna, the technology is very reproducible and is well suited to volume production.
Theoretical Antenna Radiation Performance
Figure 2 shows an array of identical individual radiating elements. The elements are spaced by a distance d, and the wave amplitude and phase at the nth radiating element are defined as An and Φn, respectively. For this discussion, the antenna will be regarded as a radiator. In a typical case, where all elements of the array are linear, the antenna will be reciprocal. That is to say that its characteristics will be identical whether it is functioning as a receiver or a transmitter.
The array factor for this array is given by the following expression
Array Factor =
λ = wavelength
Equation 1 is for a linear case, suitable for this treatment which is concerned with the azimuth direction only, when increasing the antenna beamwidth. The control of phase to the vertical array can be considered independently and is not the subject of this article.
The theoretical radiation pattern of the array of elements in azimuth is thus determined by two factors, the natural radiation pattern of the single elements and the array factor. The final radiated pattern is the product of the array pattern and the individual beam pattern. The natural beamwidth of a single patch is of the order of 60° depending upon exactly what geometry of patch is used and upon the radiation polarization desired. Figure 3 shows the theoretical radiation pattern of a single element in azimuth. This is typical for a radiating element used in vertically polarized antennas. By using this element in a suitable array, which has a tendency to direct energy away from the main lobe, it is possible to increase the beamwidth of the array to the desired value. Figures 4 and 5 show theoretical 90° and 120° nominal beamwidths produced from arrays of elements having identical radiation characteristics to that of the single element shown previously. Note that it is not easy to achieve exactly 120° beamwidth, and the actual beamwidth for a nominal 120° beamwidth antenna is generally closer to 115°.
There are other benefits that can be gained from using a flat panel array to produce broad azimuth beamwidths. It is possible to produce antennas with a variety of polarizations. Single radiating elements used to produce broad azimuth beamwidths tend to be based on dipole configurations, which only lend themselves to vertically polarized requirements. Arrays may be made of patches with a variety of polarizations, such as vertical, horizontal, slant polarization at a desired angle, or even circular polarization. The technique of using an array to broaden the beamwidth is equally applicable in any of these cases. It is thus possible to make broad beamwidth base station antennas in any polarization. More significant still, for some applications, is the ability to control the roll-off of the pattern at wider angles with an array. Many antennas have stringent requirements, not only for minimum gain values to be met across the required area of coverage, but also for the gain of the antenna to roll-off rapidly thereafter. This is needed to prevent interference with adjacent sectors, serviced from the same base station site. Many base station requirements include templates defining the maximum allowable gains outside the sectors. Within Europe, all multipoint communications systems are subject to the licensing requirements of the relevant national authorities for radio communications. These generally require the performance of antennas used in these systems to meet the specifications produced by the European Telecommunications Standards Institute (ETSI). The arrays are capable of producing patterns that meet the ETSI standards due to a rapid, off-axis, roll-off in gain, whereas a single broad beamwidth radiating element is not.
Measured Antenna Performance
Although it is clearly possible to generate theoretical radiation patterns for an array of radiators that appear to have ideal characteristics, it is altogether a different task to achieve comparable performance from a real antenna. In addition to modelling the predicted radiation pattern from an array of radiators, a detailed finite element analysis of one tier of the array used to form a complete base station was performed. Figure 6 shows the measured data from the single tier array with the modelled data superimposed. The difference in peak gain is 0.4 dB between the two cases, and whereas theory predicts a 3 dB beamwidth of 114°, the actual beamwidth was 110°. The peak gain is lower than for a normal antenna as this was only for a single tier of a larger array.
Complete antennas employing arrays in both azimuth and elevation were also built, some with nominal 90° beamwidths and some with nominal 120° beamwidths. Figure 7 shows the azimuth pattern achieved with this kind of antenna. These were achieved with an elevation beamwidth of 10° and no elevation tilt angle, but it is possible to make such antennas with different elevation beamwidths and tilt angles as required, and vary the peak gain accordingly.
Figure 8 shows a complete nominal 120° sector antenna and Figure 9 demonstrates the azimuth radiation pattern for the product. Again this antenna has a 10° elevation beamwidth. Note the gain is lower on account of the broader azimuth beamwidth. Typical operating bandwidth for these antennas is 9 percent.
Base station antennas have been built using flat panel arrays to produce wide angle beamwidths up to 120°. These have been built at a number of different frequencies with various polarizations. They have demonstrated that wide beamwidths can be achieved from an array, wider than the intrinsic beamwidth of the radiating element used. This enables antennas to be built having superior beam shape, being able to meet more difficult specifications. The use of flat panel technology results in a low profile final product. As the performance is derived from etched PCBs and all features on the circuit are relatively large, etching tolerances have a small effect on the PCB performance. This makes the antenna design very reproducible, and variations between antennas within one design are minimal. This technology is therefore well suited to volume production of base station antennas.
Chris Walker graduated with a bachelor's degree in physics from Cambridge University in 1982. He has worked with EEV, Litton and CPI in developing vacuum tubes for radar and communications systems. Other development work included developing piezoelectric tuning mechanisms, long life cathode technology and miniaturized magnet packages for magnetrons. He has been awarded five patents relating to aspects of magnetron design. He joined European Antennas Ltd. in 1998 and currently serves as the company's technical director.