In many realistic situations, however, the reflector's feed structure cannot be approximated as a simple point source or it may even cause blocking effects. In problems like this, a full-wave simulation of the entire structure is needed. , e.g. using an FDTD solver. But the enormous computational size of problems of this kind easily limits the practicality of using a full-wave solver.
Â
Most of [[EM.Cube]]'s computational modules provide a simulation observable called "Huygens Surface". This observable records the equivalent surface electric ('''J''') and magnetic ('''M''') currents (which are related to the tangential electric and magnetic field components) on the surface of a virtual box in the computational domain. According to the electromagnetic equivalence theorem, the equivalent surface currents '''J''' and '''M''' fully and accurately represent the fields produced by one of more radiators that completely circumscribed by the Huygens box. As a result, the Huygens surface observable produced at the end of a full-wave simulation like FDTD can be used as a "Huygens Source" to excite another structure possibly in a different computational module. Applying this concept to the large reflector antenna problem, we can isolate and enclose the feed structure in a Huygens box and solve it using [[EM.Tempo]]. The Huygens surface data can then be imported as a Huygens source to [[EM.Illumina]] to illuminate a large parabolic reflector.
In this application note, first we consider a parabolic reflector of manageable size illuminated by a waveguide-fed pyramidal horn antenna. We first solve the reflector-horn combination problem entirely using [[EM.Tempo]] to establish a baseline for verification. Next, the horn structure is taken out and enclosed in a Huygens box. A Huygens surface observable is defined and computed for the [[EM.Tempo]] project. The parabolic reflector antenna is then transferred to [[EM.Illumina]] and it is illuminated by a Huygens source based on the previously simulated data. [[EM.Illumina]]'s IPO solver is used to solve this hybrid problem. The results are then compared to the baseline FDTD results.