Several methods have been studied to increase the capacity of communication services, expand frequency resources, develop advanced communication services such as the 4G and Beyond 4G and 5G techniques –, and shorten the distance between the antenna and user to reduce the path loss. Although various advanced technologies such as active array antennas –, MIMO –, and beam-forming techniques – have been developed, they have not been able to keep up with the increasing speed of mobile traffic capacity.
Therefore, the next-generation mobile communication base station (BS) system separates the radio frequency unit (RFU) from the base band unit and integrates the BS antenna and RFU to minimize the cable loss between the BS antenna and the system, thus minimizing the power loss. It is important to develop a functionally distributed base station that is applicable to a variety of cell environments (indoor cells, small cells, macro cells, and so on) using a module-type (BS module) BS antenna and RFU. A BS module has the advantages of easy installation and repair, small power loss, energy savings, and simple implementation of the next-generation mobile communication BS system.
The current mobile communication BS system has the disadvantage of high power consumption because it uses a passive element array antenna structure – with 10% to 20% efficiency. Existing array antennas use a 3-dB power divider to distribute the input power to each antenna element. As each power divider has a constant loss, the antenna efficiency decreases. Higher power is used to compensate for these reduced efficiencies. In addition, the maintenance costs increase when replacing the entire base station if a failure occurs in the BS antenna. However, a next-generation mobile communication BS using a BS module can be repaired by replacing just the failed BS module instead of replacing the entire system.
The core technology of a BS module is the antenna. A cube antenna – is considered for application to a variety of cell environments and is easy to replace. However, a cube antenna has certain problems that have to be overcome. First, when the BS antenna is enveloped in a metal-type cube, the antenna boundary conditions are changed, and the antenna characteristics (bandwidth and antenna gain) deteriorate. Second, a small-sized base station is necessary for application to a variety of cells. The BS antenna size will determine the size of the entire base station. It is very difficult to design a small-sized dual-polarized antenna for a limited metal cube space.
In this paper, we present the design of a small dual-polarized BS antenna with a simple isolator applicable to various cell environments for next-generation mobile communication services. It is not possible to place a conventional dual-polarized BS antenna in a metal cube owing to its large size and because of its complex structure that ensures high isolation –. High isolation is achieved using a simple isolation patch. The proposed BS antenna has a small size, dual polarization, wide bandwidth, high isolation, and high antenna gain despite being enveloped in a metal cube. We improved the impedance matching and bandwidth using a coupled feed method.
The BS antenna is operated at 1.72 GHz to 1.89 GHz, and the measured isolation is greater than 30 dB at the operating frequency band. The measured antenna peak gain is 8 dBi. We propose improved methods for the isolation between the two BS modules for the MIMO base station system. The simulated isolation is over 50 dB between the two BS antennas.
We designed and implemented a small dual-polarized BS antenna, as shown in Fig. 1. The size of the metal cube is 100 mm × 100 mm × 30 mm, and the antenna element size is 80 mm × 80 mm × 15 mm. The microstrip (MS) coupled feed lines and simple isolation patch are placed below the MS radiation patch. The small BS antenna consists of a bottom PCB (Taconic TLY-5A, εr: 2.17), MS coupled feed lines, two metallic feeding posts (4.5 mm × 4.5 mm × 15 mm), a shorted metallic isolation post, a top PCB (FR-4, εr: 4.4), and a metal cube. The MS rectangular radiator is placed on the top plane, and the MS coupled feed lines and simple isolation patch are placed on the bottom plane of the top PCB. The MS coupled feed lines and metallic feeding posts are combined with the microstrip line printed on the bottom of the PCB. MS feeding lines were designed to have a 50 Ω impedance. The simple isolation patch is positioned at the bottom plane of the top PCB and is connected to the substrate ground plane using a metallic isolation post.
The measured reflection coefficient and isolation of the BS antenna, with and without the metal cube, are shown in Fig. 2(a). The proposed BS antenna was tested using an Anritsu Vector Network Analyzer (MS46122A) in an anechoic chamber. The BS antenna without a metallic cube was operated at 1.9 GH2 to 2.2 GH2 and did not satisfy the operating frequency band. Figure 2(b) shows the reflection coefficient and isolation with the metal cube. The proposed BS antenna was experimented on with and without a simple isolation patch. As shown in the figure, the proposed antenna was tested with and without a simple isolation patch. The measured frequency is 1.72 MHz to 1.89 MHz (port-1 and -2) with and without a simple isolation patch, respectively. When a simple isolation patch is used, the proposed antenna has a high isolation value.
The measured port-to-port isolation of the small BS antenna is greater than 30 dB over the operating frequency band (max = 45 dB). The simulations were implemented in perfect free space (vacuum), and no loss of measured cable or manufacturing losses were considered. However, the prototype antenna was operated in real free space, and there were production and measured cable losses. Therefore, the simulation and measurement results differed slightly. However, the simulated results are in reasonable agreement with the measured results.
The measured and simulated radiation patterns of the BS antenna with a metal cube are shown in Fig. 3. As shown in the figure, both the cross polarization and back lobe are suppressed using a metal cube. Measured BS antenna peak gains of 8.0 dBi at 1.735 GHz, 7.9 dBi at 1.755 GHz, 7.8 dBi at 1.83 GHz, and 7.8 dBi at 1.85 GHz were obtained. The measured 3-dB beamwidths in the yz- and xz-planes were 100º at all operating frequencies.
We simulated the effects of the simple isolator with an excited E-field distribution at 1.83 GHz. The E-field distribution was simulated using ANSYS HFSS. As shown in Fig. 4(a), when not using the simple isolator, the E-field excited at port-1 affects port-2. When using the simple isolator, the E-field is mostly focused at the isolator. As shown in Fig. 2(b), the results confirm high isolation characteristics with the simple isolator.
A small cell base station is composed of a power supply part, an IP based module, and a BS module, as shown in Fig. 5. The BS module is composed of a dual-polarized BS antenna and an RF module. The dual-polarized BS antenna is located on the top plane, and the RF module is integrated at the bottom plane.
The RF module consists of a front-end part, a PA, and a low-noise amplifier (LNA). The front-end part consists of a transmitter-receiver RF filter, a circulator, and an LNA. As shown in Fig. 5, the BS antennas are located in close proximity to each other. A signal excited from each BS antenna will significantly decrease the performance of the transmitter and receiver. It is important to suppress the leakage wave to improve the isolation between the BS antennas.
In particular, a strong leakage wave is transmitted in the feeding portion. As shown in Fig. 6, the BS antennas are arranged with different feeding directions, and the best isolation is chosen. We conducted a simulation with four vertical arrangements. The isolation values were compared to select a high isolation position. The distance between the BS antennas was lamda at 1.83 GHz.
Figure 7 shows the simulated isolations for the four cases. We conducted simulations for S31, S41, S32, and S42 to confirm the isolation between the BS antennas. The isolation values were simulated to be over 30 dB at all positions. The BS antenna position for case-2 showed a high isolation value, and thus, case-2 was chosen.
Figure 8 shows the simulated and calculated results when the case-2 arrangement is used. We applied an aluminum supporting plate to the bottom plane of the BS antennas. We improved the isolation by placing a metal isolator stub between the BS antennas. Figure 8(b) shows the values of the simulated isolation when using a metal isolator stub. Without the metal isolator stub, as shown in Fig. 7(b), the average isolation value is about 38 dB at S31, S41, S32, and S42. However, with the metal isolator stub, the average isolation value is approximately 50 dB at S31, S41, S32, and S42.
A correlation coefficient was used to confirm the interference between the antennas. An acceptable correlation coefficient for a typical MIMO system is less than 0.5 . In this study, the correlation coefficient was calculated using the scattering parameter ,  because this method provides sufficiently accurate results and a simple calculation.
Figure 8(c) shows the calculated correlation coefficient using ANSYS HFSS and MATLAB. The correlation coefficients are less than 0.05, and the two BS antennas are highly decorrelated.
A small dual-polarized BS antenna integrated RF module applicable to various cell environments for next-generation mobile communication services was designed and tested. We tested the proposed BS antenna with and without a simple isolator. When using the simple isolator, the measured isolation value was greater than 30 dB. We located a metal isolator stub between two BS antennas, and the isolation between the two BS antennas was substantially improved. A wide bandwidth, high isolation, low cross-polarization, high gain, and high decorrelation were achieved.