Search

A novel base station antenna (BSA) configuration is presented to mitigate degrading physical antenna effects in massive multiple-input multiple-output (MIMO) systems, while minimizing implementation complexities. Instead of using a commonly considered single antenna panel comprising of many elements covering a wide field-of-view (FOV) of 120 degrees, $L$ tilted panels are used employing $L$ times fewer elements and $L$ times smaller FOV per panel. The spatial resolution of each panel is enhanced by employing sparse arrays with suppressed (grating-lobe) radiation outside its corresponding FOV. Therefore, more directive antenna elements can be deployed in each panel to compensate for the effective isotropic radiated power (EIRP) reduction. While sectorisation reduces the antenna gain variation in 120 degrees FOV, cooperation among multiple panels in downlink beamforming is seen to be capable of inter-panel interference suppression for sum-rate enhancement. A network model is used as a multi-user (MU) MIMO simulator incorporating both antenna and channel effects. It is shown that when the number of base station antennas is ten times the number of users, the average downlink sum-rate in pure line-of-sight (LOS), rich and poor multipath environments is increased up to 60.2%, 23% and 11.1%, respectively, by multi-panel sparse arrays applying zero-forcing (ZF) precoding.

With the emergence of the mobile fifth generation (5G), additional challenges are introduced for the radio access antennas at the base stations that operate at mm-waves. The base station antenna is then required to deliver high EIRP to compensate for the high path loss at mm-wave frequencies. Furthermore, the massive-MIMO concept is aimed to be applied to increase the overall system capacity. For these reasons, assuming traditional phased arrays, the number of antenna elements is expected to be large which will lead to high power consumption and the need for active cooling. In this presentation, the PAF reflector antenna concept will be proposed aiming at utilizing the reflector’s high gain to produce high EIRP. That can be achieved by utilizing a small number of active elements in the focal plane thanks to the reflector antenna’s focusing properties. Moreover, multibeam generation and steering is provided by exciting different antenna elements in the reflectors focal plane. This way, high EIRP can be generated while reducing the power consumption. At the symposium, a designed PAF reflector will be presented along with relevant simulation results of the past and ongoing research.

The continued demand for high-speed wireless communications is driving the development of integrated high-power transmitters at millimeter wave (mm-Wave) frequencies. Si-based technologies allow achieving a high level of integration but usually provide insufficient generated RF power to compensate for the increased propagation and material losses at mm-Wave bands due to the relatively low breakdown voltage of their devices. This problem can be reduced significantly if one could combine the power of multiple active devices on each antenna element. However, conventional on-chip power combining networks have inherently high insertion losses reducing transmitter efficiency and limiting its maximum achievable output power. This work presents a non-conventional design approach for mm-Wave Si-based Watt-level power amplifiers that is based on novel power-combining architecture, where an array of parallel custom PA-cells suited on the same chip is interfaced to a single substrate integrated waveguide (to be a part of an antenna element). This allows one to directly excite TEm0 waveguide modes with high power through spatial power combining functionality, obviating the need for intermediate and potentially lossy on-chip power combiners. The proposed solution offers wide impedance bandwidth (50%) and low insertion losses (0.4 dB), which are virtually independent from the number of interfaced PA-cells. The work evaluates the scalability bounds of the architecture as well as discusses the critical effects of coupled non-identical PA-cells, which are efficiently reduced by employing on-chip isolation load resistors. The proposed architecture has been demonstrated through an example of the combined PA with four differential cascode PA-cells suited on the same chip, which is flip-chip interconnected to the combiner placed on a laminate. This design is implemented in a 0.25 um SiGe BiCMOS technology. The PA-cell has a wideband performance (38.6%) with both high peak efficiency (30%) and high saturated output power (24.9 dBm), which is the highest reported output power level obtained without the use of circuit-level power combining in Si-based technologies at Ka-band. In order to achieve the optimal system-level performance of the combined PA, an EM-circuit-thermal optimization flow has been proposed, which accounts for various multiphysics effects occurring in the joint structure. The final PA achieves the peak PAE of 26.7% in combination with 30.8 dBm maximum saturated output power, which is the highest achievable output power in practical applications, where the 50-Ohms load is placed on a laminate. The high efficiency (>20%) and output power (>29.8 dBm) over a wide frequency range (30%) exceed the state-of-the-art in Si-based PAs.