EM Simulation Software Efficiency Demonstrations
Radomes impact the performance of the antennas that they protect. This is especially true of AESA antennas. There is a trade-off between providing protection from the environment to the elements and enhancing the performance of the overall system. It is often challenging or impossible to perform a detailed analysis before fabricating such a complex system.
In many cases, the design tools used to design the radome and antenna structure are entirely different and this disaggregated approach can lead to additional costs and schedule risk. Without the ability to fully assess the combined performance of the antenna and radome in a simulation, there is the chance that design issues may not be apparent until later in the fabrication, assembly and qualification process. The further through the process that any issues are identified; the higher costs of mitigation.
Figure 6 Cost increase of design flaws at various stages of the production cycle.
Figure 6 illustrates this concept. A key stage in the entire process of developing an active antenna system is the calibration of the boresight stage. This is generally the final stage in the process once the radome and antenna system are fully assembled and characterized. Traditional EM simulation tools make it challenging, if not impossible, to analyze the antenna and radome as a system. With appropriate EM simulation resources, the entire antenna system and radome could be fully analyzed with weights optimization, uncertainty analysis and perhaps even simulation-based calibration, which would be a benefit.
Radome & Antenna Array Example
Figure 7a shows a 30 x 15 dipole triangular lattice array with a curved quartz-cyanate ester (QCE)/foam/QCE multilayer radome material. Figure 7b shows an example of the 3D antenna pattern from this antenna when it is simulated in Nullspace EM. Beam steering error analysis is a critical tool to minimize this error in the final design. This type of analysis is generally done with a drastically simplified version of a simulation or after fabrication.
Figure 7 (a) A 30 x 15 dipole triangular lattice/radome model. (b) The 30 x 15 dipole triangular lattice 3D antenna pattern simulated in Nullspace EM.
Figure 8 (a) E-plane steering angle error for the 30 x 15 dipole triangular lattice antenna. (b) H-plane steering angle error for the 30 x 15 dipole triangular lattice antenna.
The Nullspace EM solver changes this process. In this example, the Nullspace EM software simulated all the beam steering angles for the entire antenna system and radome combination and this simulation ran roughly 7x faster than a leading commercial EM simulation tool that considered only one beam steering angle. A direct comparison of all scan angles was unrealizable between the EM software packages, given the time and computational resources required by the leading EM simulation tool.
Figure 8a shows the EM simulation results in the E-plane and Figure 8b shows the H-plane plot. It is possible to perform a parametric analysis and determine the contributing factors to the beam steering error. A sensitivity analysis can be performed to varying tolerances to understand the stack-up of the beam steering error. This is only possible with the detail allowed by more efficient and modern EM simulation software. The design resource headroom afforded by using more efficient EM simulation software can be instrumental in predicting practical performance or driving the requirements for the various stages of fabrication to ensure that production devices meet the desired performance and quality standards.
Patch Antenna Simulation Accuracy Experiment
To analyze the potential accuracy discrepancies with legacy EM simulation software more fully, an experiment involving patch antennas was devised. Figure 9a shows an example of a relatively small patch antenna designed to operate in the GHz range. To remove fabrication variability considerations, 12 patch antennas were fabricated in two batches, with six each from two different vendors. The patches were also physically measured using a profilometer to ensure the fabrication accuracy of the patches. The patch design is indicative of the type of patch antenna commonly used in antenna array technologies. Figure 9b shows the profilometer results for the patch antenna to ensure the fabrication tolerances compared to the 3D model used in the simulation.
Figure 9 (a) Sample patch antenna. (b) Profilometer results for the patch antenna.
Figure 10 S11 versus frequency simulated and measured results for the patch antenna samples.
The patch designs were also simulated with a leading commercially available EM simulation software, as well as Nullspace EM.3 The plot in Figure 10 shows the minimum and maximum S11 response of all 12 patch antennas predicted by the leading software, along with Nullspace EM’s results. The results show that Nullspace’s EM simulation software is more accurate out of the box than the other commercially available EM software suite. Typically, the solution in these cases is to tune the EM simulation software to yield better agreement with the given design and fabrication process. While this will improve the accuracy, it also creates an iterative loop to achieve the necessary levels of accuracy and trust in the simulation software when modeling new designs. The Nullspace EM software avoids this iterative step.
Conclusion
EM simulation has become an essential tool for RF product design and technology development and many stages of the design cycle depend largely on the accuracy and efficiency of the EM simulation software. Greater levels of EM simulation accuracy and efficiency can have significant impacts on the overall design cycle time and the number of iterations a design team must go through to meet with success. Recent advances in EM simulation technology optimized for complex and electrically large simulations can now deliver many times the speed of legacy EM simulation codes while maintaining high accuracy.
References
- M. A. Horn, T. N. Killian and D. L. Faircloth, “Multi-GPU Accelerated Fast ACA Direct Solve for Moment Method Solution of Composite Bodies,” 2014 IEEE Antennas and Propagation Society International Symposium (APSURSI), 2014, pp. 1630-1631, doi: 10.1109/APS.2014.6905141.
- Z. Cendes, “The Development of HFSS,” 2016 USNC-URSI Radio Science Meeting, 2016, pp. 39-40, doi: 10.1109/USNC-URSI.2016.7588501.
- Nullspace EM, Web: www.nullspaceinc.com/nullspace-em.
- T. Killian, D. L. Faircloth and S. M. Rao, “Acceleration of TM Cylinder EFIE with CUDA,” 2009 IEEE Antennas and Propagation Society International Symposium, 2009, pp. 1-4, doi: 10.1109/APS.2009.5171648.
