Software Platforms Bridge the Design/Verification Gap for 5G Communications Design - Part II

Oct 26, 2016 | Posted by: David Vye

View Part 1 >>

The integration of simulation technologies, system prototyping tools, and automated test equipment by NI AWR Design Environment and LabVIEW platforms is critical for addressing the complexity of developing 5G wireless technology.  In these cases, design teams will need to rely on a combination of simulation and prototype testing in order to ensure design robustness.  Although simulation is essential to design a test bed or prototype, measurement is often needed to validate assumption made before fabrication, and to convince others of design viability.

For instance, Gary Xu, director of research at Samsung, showcased one of the first public demonstrations of a prototype for a 5G full-dimensional MIMO (FD-MIMO) base station at NIWeek 2015 in Austin, TX. The demo comprised a small-base station containing the FD-MIMO antenna array and four NI USRP RIO software receivers, which emulated into four “5G” handled terminals. One of the key technologies of the demonstration was Samsung’s new 3D beamforming algorithms. The prototype demonstrated that 3D beamforming enabled higher throughput and an increase in the number of supported users.  In this instance, the system improved from enabling 2 Mbps for one user to 25 Mbps for four users with the use of 3D beamforming.  Dr Xu’s demonstration, shown in Figure 1, is an excellent example how how radio prototyping is an important tool for proving the viability of a given system design.

Figure 1: Samsung FD-MIMO 3D beamforming prototyping system implemented with four NI USRP RIO software receivers.

On the simulation side, NI AWR Design Environment helps engineers develop radio components such as PAs, filters, antennas, sources such as voltage-controlled oscillators (VCOs), control circuits such as mixers, converters, and switches, and interconnect technologies before fabrication. For example, 5G will require significant use of advanced antenna technologies.  Because of the high-bandwidth, low-power requirements and considerable channel losses at mmWave frequencies, antenna designs will be needed with multiple, directed beams and polarization diversity and control (as in Figure 1 above).  Phased arrays are an obvious candidate to meet these requirements.  This in turn requires the circuit simulation and system environments to support phased array simulations.

Traditionally, the phased array is simulated in an EM simulator and the resulting S-parameter file is embedded into the circuit simulator to complete the design.  The integrated circuit and EM simulators allow the designer to investigate how the antenna and circuit interact with each other.  In particular, the impedance of the array’s ports changes with the scan angle of the beam.  In turn, the performance of the PA driving the antenna is severely affected by the port impedances.  The designer must often go back and forth between the antenna simulation and circuit simulation to accurately model this behavior.  Microwave Office software can now automatically couple the two simulations.  The PA “sees” the changing port impedance and the antenna scans its beam as the input power and phasing to the input ports is changed.  Along with saving time and reducing errors, the designer can now optimize and perform yield analysis on circuit/antenna systems. Figure 2 shows how EM co-simulation will continue to play an important role in the development of densely populated high-frequency electronics and interconnect characterization.

Figure 2: (a) patch antenna array design using AXIEM for planar EM simulation and (b) system-level simulation of a phased array.

Recent work from Université du Québec à Rimouski (UQAR) students led by Dr. Chan-Wang Park provides an example of simulation and test data being applied together in a 5G design. The design team developed a 6 watt, 1 GHz PA for use with 5G MIMO multi-carrier signals. Because they wanted to linearize the PA in the future, the team intended to correct the nonlinearity of the PA by using a neural network pre-distortion linearizer, Volterra, or polynomial pre-distortion linearizer developed with NI AWR Design Environment and LabVIEW. A digital pre distorter in the baseband will be used to create an expanding nonlinearity that is complementary to the compressing characteristic of the PA.

The design team was able to achieve first-pass PA design success through detailed circuit/EM co-simulation using Microwave Office and AXIEM, along with scalable high-frequency Modelithics models of the Cree GaN transistor, imported multi-harmonic source and load-pull data (for impedance matching and model validation) from Focus Microwave, and the NI PXI RF instruments automated with LabVIEW for fast test results of the PCB-based prototype, shown in Figure 3.  The team was able to develop a simplified solution for future 5G MIMO telecommunication system standards using a pre-distortion linearizer, which will be implemented in NI LabVIEW software and executed on a Xilinx Virtex-6 FPGA.

Figure 3: (a) Load/source pull data and test set-up diagram used for model verification and impedance matching (fundamental and 3rd harmonic) and (b) PXIe-based RF measurement platform for 4x4 MIMO.

Together these interoperable platforms will give design teams the power and flexibility to realize the high-frequency electronics that will drive 5G and IoT.