Excitations for SBR+

Antenna and SBR+ Hybrid Region with Hybrid and Arrays Solution Type

Within the Hybrid and Arrays Driven Modal and Driven Terminal solution types, HFSS can use SBR+ to compute the scattered field contribution from a platform geometry. More information can be found in Assigning SBR+ Hybrid Regions.

When an antenna is simulated via FEM, the antenna must be enclosed by a Radiation boundary or FE-BI Hybrid Region. Instructions may be found in Assigning Radiation Boundary and Assigning FE-BI Hybrid Region. First, the antenna is solved by HFSS in the absence of the SBR region. The bounding surface of the Radiation Boundary or FE-BI Hybrid Region is also used as the exterior Huygens surface whose field samples (equivalent currents) yield current sources or far-field radiation patterns to drive the SBR+ simulation or serve (via reciprocity) as a receiving antenna in an SBR+ simulation. When Near Field is selected for Field Type, currents are computed on the Huygens surface using

J = nXH,

M = -n X E,

where J and M are the equivalent electric and magnetic current sources, n is the outward normal of the Huygens surface, H and E are Electric Fields and Magnetic Fields computed by FEM solver.

A Perfect E Boundary may be assigned to some faces of the exterior Huygens surface, but at least one face must be assigned to Radiation Boundary or FE-BI region. When Perfect E is assigned on a face, no currents for SBR+ simulations are generated on that face. More instructions are found in Assign a Perfect E Boundary. Also, the presence of Perfect E indicates that the antenna has a ground plane, and the antenna can be properly installed on a metallic platform surface. When Current Conformance is set to Enable, current conformance is only performed on antennas where at least one face is assigned to Perfect E. When Current Conformance is set to Auto, current conformance is only performed on antennas placed with a small tolerance separation from the nearest SBR+ region surface, this tolerance being determined as one-twentieth (1/20) of the diagonal displacement of source’s bounding box. In addition, if none of the faces is assigned a Perfect E infinite ground plane, Auto will automatically insert one on the face nearest the SBR+ region.

When SBR and IE hybrid regions are assigned to different objects, the FEM with IE regions must first be solved together via two-way coupling. The instructions are found in Set Current Source Option. When two-way coupling is configured, current sources are also generated on the IE region to launch rays in an SBR+ simulation.

Antenna introduced via Linked Source

The antenna may be introduced via linked sources for SBR+ simulations. When a designed antenna is already analyzed, a linked source is a convenient way to introduce the antenna into SBR+ simulations without analyzing the antenna again. Instructions for configuring a linked source are found in Link from a Driven Modal or Driven Terminal Antenna to an SBR+ design for SBR+ solution type, and Far Field or Near Field representations for Driven Modal or Driven Terminal solution type.

When a near field linked source is introduced for SBR+ simulation, fields and current sources are also generated on the exterior Huygens surface.

A far-field linked source is converted to a pattern-weighted point-source for the SBR+ simulation. It is therefore handled with much greater efficiency by SBR+ than a near-field linked source, which presents SBR+ with a distribution of current sources, each of which launches its own ray set. However, the far-field representation is only valid if the SBR+ scattering geometry is in far-field of the source. More discussion on this crucial point is found in the later section, Near Field vs Far Field Source Representation. Moreover, the phase reference of the far-field representation must be manually specified to be as close as possible to the physical antenna phase center by selecting Excitations>Set Far Field Phase Center... in the source design of the antenna. More instructions are found in Link from Driven Modal or Driven Terminal Antenna to SBR+ Design.

Incident Plane Wave

A plane wave is a wave that propagates along a fixed direction where the electric and magnetic fields are in the transverse plane and perpendicular to each other. For SBR+, Incident Plane Wave is required to compute Radar Cross Section (RCS). More instructions are found in Incident Plane Wave.

The phase reference of the incident plane wave may be set as the Zero Phase Position in the Incident Wave panel. For monostatic RCS simulation, this serves as the phase reference for both illumination and observation. For bistatic RCS simulation, the Far Field Radiation Sphere Setup allows one to set the phase reference for observation, which can be set up via a relative coordinate system.

Parametric Antennas

The SBR+ Solution Type provides a suite of parametric antennas that are very convenient for system analyses. Parametric antennas are sources that either have no physical structure definition or have a highly idealized physical structure. No model parts are constructed or used in the simulation. Parametric antennas are intended to be very quick to define and place, copy and duplicate, and are thus intended for systems studies where the specific antenna structure is either unknown or where a simple, topologically representative antenna type is sufficient. The editable parameters depend on the antenna type. More detailed instructions are found in Parametric Antennas for SBR+ Solution Type.

Parametric Antennas create far-field pattern sources that can provide S-parameter outputs and composite far-field radiation patterns. Some Parametric antennas also provide a current-source representation option. In general, a current-source representation yields more accurate results than its far-field counterparts, especially when placed nearby platform scattering geometry within 2 or 3 wavelengths. This is because it provides a near-field description of the antenna that is physical everywhere, whereas the far-field description is only physical beyond a critical distance from the antenna where it’s far-field has converged. In particular, current-source antennas with implicit ground-planes (e.g., Wire Monopole, Parametric Slot, Horizontal Dipole, and Cross-Dipole over Ground Plane) are intended to be placed directly on a highly conductive surface of the platform (SBR+ scattering region). In these situations, using the same antenna’s far-field counterpart is not recommended. For example, if you place the far-field version of the Wire Monopole on a supporting platform that provides the grounding function to the wire monopole, the half-donut far-field pattern of the Wire Monopole will only interact with structures in the upper half-space of the placement surface. In contrast, when using the current-source representation of the Wire Monopole, the axial current sources radiate in all directions, and SBR can correctly predict the interaction with the placement structure directly underneath because of the higher-order fields imbedded in the current-source representation.

By File Antenna

User defined far-field data in .ffd format may be imported into the SBR+ solution type. The SBR+ solver will automatically interpolate the source fields from the file-based antenna in order to represent the antenna at all frequencies of the user defined frequency sweep in the solution setup. The file-based antenna must be defined with at least two frequencies, and the min/max frequencies of the simulation sweep should fall within the frequency band of the file-based antenna. In the model window, the antennas display as a scaled far-field pattern at the antenna origin. Instructions are found in Parametric Antennas for SBR+ Solution Type. The file formats for .ffd are found in Frequency Independent Far Fields or Frequency Dependent Far Fields.

Current Conformance

The instructions for activating current conformance are found in Set SBR+ Source Region. It is only available for antennas represented by Near Field and parametric antennas that allow an Antenna Representation set to Current Sources. This capability is primarily intended for antennas requiring a ground plane to operate (e.g., monopole, blade, microstrip patch, slot, etc.) and that will be placed on a flat or slightly curved metallic section of the platform that conceptually provides this grounding function. It is typically used to make small positional adjustments to current-source positions that will improve simulation accuracy. There are a couple of reasons why one should consider using the current conformance feature.

• The unadjusted sources may otherwise be too close or possibly even slightly underneath the platform placement surface due to small positioning errors in placing the antenna on the platform. In general, placing sources that are conceptually intended to be directly on the surface closer than λ/40 above the surface harms run time with little or no accuracy benefit, as in the case of horizontal magnetic sources representing a slot cut in the placement surface.
• The platform placement surface is slightly curved, the shape of the actual antenna will conform to this curvature, yet the antenna might not exhibit this intended curvature (it is flat).

These situations are fairly common, so the current conformance feature is a valuable capability. When the antenna structure is to be placed directly on the platform, such that it touches the platform, conformance is often beneficial.

Current Reduction

For SBR+ simulations where the Tx and/or Rx antennas are defined by an equivalent currents on a large Huygens surface (e.g., FEBI sources), the current-source reduction feature can significantly improve run time by removing a large number of very weak sources. The instructions for setting up the current reduction are given in Set Current Source Option.

When antennas are represented by current sources, run time generally grows with the number of current sources. When applied to both the Tx and Rx antenna in an antenna coupling simulation, this effect is multiplicative, not additive. Antennas are sometimes represented by hundreds or thousands of current sources, and this can result in excessive SBR+ run times. The SBR+ solver will dynamically aggregate current sources when illuminating far-away surfaces to mitigate this run-time impact. While quite effective, it only applies to Tx antennas, not Rx antennas. At the same time, many of the current sources may be very weak, with the dominant ones just a small fraction of the total. One way to reduce run time without significant compromise of accuracy is to thin the current sources according to their relative powers, and this is the function of the Current Reduction option. Current reduction (optional) and dynamic aggregation (always on) can operate simultaneously, and current reduction has the further advantage of also speeding up computations involving Rx antennas.

When Thin Sources is enabled, for each SBR+ source defined by current sources, HFSS will first remove the weakest current sources during a pre-processing stage of the SBR+ region analysis. The Power Fraction to Keep can be set between 0 and 1, but settings less than 0.9 are not recommended due to the potential for excessive lost accuracy. During pre-processing, the current sources of each SBR+ source are sorted from weakest to strongest on a power scale. Their individual power contributions are summed to establish a reference composite power. If, for example, the Power Fraction to Keep is set to 0.95, then the partial sum of the strongest current sources that contribute 95% of this reference level are retained, and all the weaker current sources that contribute the remaining 5% are discarded. For sources defined by a large number of current sources, it will typically be the case that the fraction discarded will far exceed their fractional power contribution. This is why current-source thinning can be an effective means of reducing SBR+ simulation burden without significant sacrifice of accuracy.

One side effect of current-source thinning in HFSS Hybrid designs (but not in the SBR+ solution type) is that field and S-parameter results may show variability depending on HPC settings and how frequency samples of the analysis are dynamically divided among several processes to be solved in parallel. To achieve maximum efficiency, the weakest-to-strongest sorting by the thinning algorithm considers the power weight of each source averaged across all analysis frequencies. In the context of an HPC parallelization by frequency decomposition, this average can only be performed for the subset frequency samples known to each process. For example, if the analysis is divided into two separate processes, the frequency decomposition may assign N1 frequencies to the first process and N2 to the second, where N1 + N2 is the total number of frequencies configured for the analysis. The first process will be unaware of the N2 frequencies of the second process, and the current-source weights for the second process may not yet even be determined by the time the first process proceeds with the SBR+ stage of its analysis. In general, owing to the frequency dependence of the current source weights, the weakest current sources discarded by the first process will be different than the those discarded by the second process, which in turn will be different from the weakest current sources discarded if all N1 + N2 frequencies had been considered together. This impact is typically small in terms of final results, but to the extent that consistency is paramount, one should either avoid Current Reduction or switch to using a near-field link in the SBR+ Solution Type; the SBR+ Solution Type relies exclusively on ray-based decomposition for parallelization and does not use frequency decomposition.

Near Field vs Far Field Source Representation

Representing an antenna by its near-fields on a Huygens surface is always more accurate than representing it by a far-field pattern. This is because the current distribution produced on the Huygens surface captures both the near-field and far-field radiation fields of the antenna, whereas the far-field pattern sacrifices near-field information. By reciprocity, the same is true when an antenna is used as a receiver in antenna coupling simulations.

If there are surfaces in the SBR+ scattering geometry in the near-field of the antenna, then it is not suitable to represent the antenna in SBR+ by its far-field antenna pattern. Thus, when representing an antenna directly mounted on a grounding surface, such as a fuselage, the antenna should always be specified according to its current distribution, not its far-field pattern. Otherwise, one cannot expect accurate results, especially in the shadow region under the mounting surface.

As a practical matter, installed antennas always require some sort of mounting structure, and since the part of this structure attaching to the antenna must be in its near-field, using the far-field antenna pattern representation might never seem justified. This would be the wrong conclusion. For example, consider the case of a reflector antenna with a feed horn at its focal point. Suppose we wish to model this case in SBR+, where the horn is the source antenna and the reflector is the SBR+ scattering geometry described by the CAD model. The feed horn is attached to a mounting plate that is supported by thin rods extending from the outer rim of the reflector and converging on the sides of the mounting plate. As a very good approximation, we can model this by specifying in SBR+ the far-field pattern of the feed horn. Now, technically speaking, the sections of supporting rods near the mounting plate are in the near-field of the horn. However, depending on the details, these are not significant scatterers in the context of determining the radiation pattern of the reflector.

We may even choose to include the supporting rods in the CAD model, recognizing that the scattering from the rod sections near the horn is not accurately captured since they are illuminated by a far-field pattern that has not yet had sufficient radius to form. However, since scattering from these nearby portions of the support rods is small compared to the overall scattering of the reflector, this problem can possibly be ignored.

Another example would be an input antenna mounted on the front of a wing pod. If the antenna is self-contained and predominantly radiates in the forward direction (such as a cavity-backed spiral), a reasonable approach is to model this antenna according to its forward half-space antenna pattern (or perhaps an even larger conical section), setting the fields in the backward half-space (or conical section) to zero. That way, one does not illuminate the mounting pod with far-field data that are not relevant to the mounting conditions and that have, at any rate, not yet had sufficient distance to be valid. Alternatively, we may remove the mounting pod from the CAD model.

With a current-source representation, run time grows in proportion to the number of current sources in the Huygens surface, although this behavior is substantially mitigated by the SBR+dynamic aggregation feature (described below). This is because a separate set of rays must be launched from each current source and then processed. Thus, the run time for a Tx antenna described by a single current source is roughly the same as that for a Tx antenna (placed in the same environment) represented by a far-field antenna pattern. However, if the antenna is instead represented by N current sources, the run time will be up to N times as long. For this reason, if accurate results can be obtained using the far-field pattern representation, and run time is a consideration, this is the approach that should actually be used.

There is an important caveat to the above point about run time. Many antennas need to be characterized by a large number of current sources, numbering in the hundreds or thousands. With a brute force simulation, the SBR+ run times could then be excessive. However, SBR+ invokes an important feature known as dynamic zoning that greatly mitigates the run-time impact of having a large number of current sources with negligible loss of accuracy.

Given the above guidance, it is important to understand what is meant here by the complementary terms, "near-field" and "far-field". The far-field region of an antenna extends from some minimum distance Rmin to infinity, where R is the radius measured from the center of the antenna. That is, we require R > R_min for an observer at R to be in the far-field of the antenna. For a transmitting antenna, R_min is the distance beyond which the radiated fields decay as 1/R and are transverse (perpendicular) to the observation direction. If one of these conditions is not satisfied, then one is in the near-field of the antenna.

Owing to the continuous nature of fields, there is no precise way to quantify R_min for a particular antenna. That is, the far-field conditions identified above are only precisely satisfied in the limit as R goes to infinity. However, there are well established formulas for specifying R_min beyond which these conditions are satisfied to an excellent approximation. Specifically, we require

and

The first criterion is the parallax far-field criterion. Here, D is the largest linear dimension of the antenna. It is generally relevant for large aperture antennas and antenna arrays. The second criterion is based on the distance beyond which only the propagating field components remain and field components associated with stored energy and charge contributions have decayed to negligible levels. Here, R must be measured conservatively (i.e., measured between the observation point and the nearest point on the antenna). The second criterion is more relevant to small, single element antennas, such as a dipole or monopole. Regardless, if either of these criteria is violated, then one is in the near-field of the antenna, whether it is serving as a Tx or Rx antenna. If there are significant scattering surfaces in the CAD model within either of these distances, then a current-source distribution representation should be used.