Abstract
The paper presents a novel dual polarized dual fed Vivaldi antenna structure for 1.7–2.7 GHz cellular bands. The radiating element is designed for a base station antenna array with high antenna performance criteria. One radiating element contains two parallel dual fed Vivaldi antennas for one polarization with 65 mm separation. Both Vivaldi antennas for one polarization are excited symmetrically. This means that the amplitudes for both antennas are equal, and the phase difference is zero. The orthogonal polarization is implemented in the same way. The dual polarized dual fed Vivaldi is positioned 15 mm ahead from the reflector to improve directivity. The antenna is designed for 14 dB impedance bandwidth (1.7–2.7 GHz) with better than 25 dB isolation between the antenna ports. The measured total efficiency is better than 0.625 dB (87%) and the antenna presents a flat, approximately 8.5 dB, gain in the direction of boresight over the operating bandwidth whose characteristics promote it among the best antennas in the field. Additionally, the measured cross polarization discrimination (XPD) is between 15 and 30 dB and the 3 dB beamwidth varies between 68° and 75° depending on the studied frequency.
1. Introduction
Recently an active antenna system (AAS) has been proposed as a technology concept for cellular base stations [1] and it has received much attention from the industrial perspective [2–4]. Traditionally, in base station antenna arrays, the radiating elements are linearly oriented along a vertical line, and the array is fed with the fixed power ratio and relative phase with a limited number of beams in the elevation angle. In the AAS, the radiation pattern is dynamically and electrically adjusted in the elevation and azimuth angels. The AAS is interesting from the mobile communications point of view as it offers capacity enhancement, improved network availability, and especially higher energy efficiency for the future wireless communications demands with 3D beam forming capability [5].
Active antennas can create and adjust beams inside the mobile base station cell by changing the relative signal phase and amplitude of every radiating element. The beam forming can be based on either constructive interference that amplifies the beam in a given direction or a destructive interference that focuses the beam precisely. These can be applied for both transmitted and received beams independently. Beam forming enables a variety of operations according to the users, different diversity techniques, carrier frequencies, radio systems and multiple cells, and even multiple operators [2–4].
The requirements and hence the technology of antennas for AAS base stations largely depends on the application considered. Outdoor mobile communications need well controlled patterns and high power handling and environmental demands like variable weather conditions. Single antenna element performance plays an important role in AAS in terms of properties such as good impedance matching, high isolation between adjacent antenna elements, low mutual coupling between orthogonal polarizations, and high total efficiency. For commercial and competition reasons these technical properties should be designed and manufactured with low costs.
Currently available vertical micro strip patch arrays [6] or dipoles [7] can be used for interleaving different frequency bands in the AAS usage. Other types of wideband antenna elements for base stations can be found in [8–11]. Polarization diversity is typically used with dual polarized antenna elements with high isolation between the polarizations [7–15]. Complete antenna arrays can be found in [16–19]. State-of-the-art comparison of currently studied antenna radiators is collected in Table 1, where the studied antenna structure can be found in column one.
The most common way to implement two orthogonal polarizations with Vivaldi antennas is to place them orthogonally along the outer edge of each element [20–22]. Another way to implement a wideband dual polarized antenna is to orientate two Vivaldi antennas into a cross-shape with respect to the antenna center where a galvanic contact is avoided by a small longitudinal gap between antenna elements [23–25]. Dual fed Vivaldi antennas do not exist.
This letter presents a dual polarized dual fed Vivaldi antenna for 1.7–2.7 GHz frequency band keeping in mind the AAS concept for mobile base stations. The antenna is designed with two parallel dual fed radiating elements for both polarizations to increase the overall aperture of the antenna, and, thus, to increase the directivity. The antenna presents the measured −14 dB impedance matching and isolation between the antenna ports better than 25 dB over the operating bandwidth with good radiating properties. The performance is compared with the existing antenna types. Section 2 describes the radiating element and the implementation, Section 3 describes simulation results, and Section 4 describes measurement results. Finally, the conclusions are given in Section 5.
2. Radiating Element
Figure 1(a) presents a dual polarized antenna structure on Rogers RO4003C laminate containing two pairs of orthogonally oriented dual fed Vivaldi radiators located on rectangular form. The separation of radiators for one polarization is 65 mm (~λ/2 wavelength) at 2.38 GHz center frequency. The tapered opening of Vivaldi is elliptical ( = 23 mm, = 33 mm) as can be seen in Figure 1. The idea of using two symmetrical Vivaldi antennas for one polarization is to increase the aperture of the radiating element.
(a)
(b)
(c)
(d)
(e)
The width of a radiator is 61 mm and height is 46 mm, and the distance from the reflector is 15 mm. The size of the radiator is 125 mm × 240 mm and it is so called floating structure fed by a symmetrical feeding network (Figure 1(b)). The size of the reflector was optimized for low ground currents and high antenna efficiency and, on the other hand, not too high relative distance to adjacent subarray at 2.7 GHz that would increase grating lopes in the antenna array. The floating reflector here means that the antenna and the reflector are not in the same ground potential. Symmetrical feeding is implemented by using a strip line to separate the feed from the reflector. As it can be noticed, from the back of both Vivaldi antennas, an area with the radius of = 10 mm of conducting material is taken out. This is done to reduce the coupling of the radiated field to the reflector.
The connection between the symmetrical feed network and Vivaldi radiator is done with a coaxial transmission line maintaining a 50 Ω characteristic impedance. The coaxial feed line also neglects the effect of feedthrough in the reflector by minimizing the coupling of the field. The measured prototype antenna is presented in Figure 1(e) with the reflector dimensions.
In the following sections, CST Microwave studio was used to perform the antenna simulations, HP 8510C VNA was used for impedance matching measurements, and Satimo Starlab was used to measure the radiation properties of the antenna.
3. Simulation Results
Simulated impedance matching of the dual polarized dual fed Vivaldi radiator is presented in Figure 2. Both polarizations (Ports 1 and 2) were simulated from 1.5 to 3 GHz and the results exhibited better than −14 dB impedance matching over the 1.7 GHz–2.7 GHz band with the mutual coupling lower than −27 dB between antennas of Ports 1 and 2. Port length of 1 and 2 is different (two-layer feed) that causes different matching response. Simulated total efficiency is shown in the same figure and it presents better than −0.2 dB (95%) over the 1.7 GHz–2.7 GHz band. This can be considered good for base station application.
The simulated theta and phi polarization components of the dual polarized dual fed Vivaldi are presented in Figure 4. The XPD results calculated from the figure are 38 dB for theta and 25 dB for phi polarization components at 1.7 GHz. At 2.7 GHz the XPD for the theta component is 18 dB and it is 28 dB for phi, respectively. The variation of XPD is 18–42 dB over the frequency range.
4. Measurement Results
Measured impedance matching of the dual polarized dual fed Vivaldi radiator is presented in Figure 3. Port 1 presents better than −13 dB impedance matching, whereas Port 2 shows better than −17 dB impedance matching over the 1.7 GHz–2.7 GHz band. The mutual coupling between the Ports 1 and 2 is below −25 dB. Discrete ports were used in simulation whereas realization was made by connectors and thus the matching is different. Additionally, the measured total efficiency was better than −0.625 dB (87%) over the band for both ports.
Measured total gain in the boresight direction (θ = 0°) is presented in Figure 5. As it can be observed, the gain is rather flat over the frequency range varying between 7.75 and 9.2 dBi. Notice that the simulated main polarization in Figure 4 is comparable to maximum gain as the antennas are linearly polarized.
The measured XPD of the dual polarized dual fed Vivaldi is presented in Figure 5. As it can be observed, the XPD varies between 15 and 30 dB and correlates well with the simulated ones.
Figure 6 presents measured radiation patterns at -cut (ϕ = 90°,θ = 0°) at 1700 MHz, 2200 MHz, and 2700 MHz. It can be observed that the 3 dB beamwidths are 70°, 68°, and 75°, respectively. Electrical results are comparable with antennas [9, 13, 15] and the antenna complexity will decide the choice order for economic manufacturing. Antenna array concepts at [16–19, 24, 25] can be straightforwardly applied with the studied structure. Presented dual polarized dual fed Vivaldi antenna has wideband, well matched, high isolated, and high total efficiency structure with flat gain response when compared to other antennas in Table 1. Antenna array performance was simulated and presented in Figure 7 as eight-element linear array such as [8]. As it can be observed, the grating lobes are arising at 2.2 GHz in ± 90 deg and at 2.7 GHz in ± 60 deg. The maximum gain in boresight is around 18 dB as sidelobe levels are around 15 dB below the maximum gain.
(a)
(b)
(c)
5. Conclusions
The dual polarized dual fed Vivaldi antenna element for 1.7–2.7 GHz frequency band was designed for AAS concept utilized at mobile base stations. Simulated and measured performance exhibited 45% relative impedance bandwidth with return loss better than −14 dB and mutual coupling between antenna ports was better than −25 dB. The antenna gain in the boresight was 7.75–9.2 dBi and the variation between the polarizations was observed to be very small. The total efficiency was −0.625 dB (87%) over the band for both ports. The half power or 3 dB beamwidth was approximately 70° over the antenna operating bandwidth. In conclusion, the antenna performance is sufficient to be utilized in an active antenna system in mobile base stations.
Conflicts of Interest
The authors declare that there are no conflicts of interest regarding the publication of this paper.
Acknowledgments
This work was supported by Finnish Funding Agency for Innovation and Nokia Networks.