AWR eBooks

Best Practices for Efficient and Effective Planar EM Simulation

Issue link: https://resources.system-analysis.cadence.com/i/1326178

Contents of this Issue

Navigation

Page 13 of 21

Best Practices for Efficient and Effective Planar EM Simulation 14 www.cadence.com/go/awr If all else fails, the HARF size limit can be increased so that the problem can be simulated. This setting is a secondary setting under the Meshing Options tab for the EM project. The default setting is usually 1000 and can be set higher so the warnings and errors don't occur. The concern with HARF elements is that they can cause the matrix to behave badly, making a solution difficult or impossible. The user will see the solver begin to struggle by requiring many iterations leading to long solve times. Ultimately, the AWR AXIEM simulator will give an error that the matrix cannot be solved. If the problem solves with the higher HARF settings, the elements in question didn't negatively affect the matrix enough to result in significant problems in the solve. It is also a good idea to check the answer with passivity and energy tests (covered in Best Practice Tip #10). The ideal mesh is one that models the currents on the metal accurately enough that a useful answer is obtained that is not overly large and will not unnecessarily slow down the simulation. The designer has some control over the mesh without resorting to extreme measures like modifying the layout. In a typical PCB or module layout, most of the facets are used for the ground planes and vias. Typically, these types of geometries have large numbers of vias and pads, especially grounding vias. Geometry simplification rules should be employed to simplify the via shapes and to merge grounding vias into fewer shapes whenever possible, as discussed in Best Practice Tip #4. Mesh properties can be changed for a single polygon by selecting the polygon in the layout, for example a ground plane, and then changing the mesh properties by selecting the Mesh tab in the Properties menu of the selected shape. Options are available for setting the mesh density, the decimation level of the mesh, and if the shape is drawn with non-zero thickness. (See Best Practice Tip #6 for more details on metal thickness.) Figure 15 shows the mesh for a ground plane with normal mesh density and low mesh density. The density of the mesh is usually set to low for ground planes. Typically, the results of a simulation do not depend critically on ground plane mesh size, as long as the plane is included in some way in the simulation. Figure 15 also illustrates the concept of mesh mirroring, the option for which is set in the meshing options – Mesh Mirroring. If enabled, any ground plane that is on a nearby layer in the STACKUP that is within the vertical distance of the mesh mirroring distance will have the same mesh as the line. In Figure 15, the mesh mirroring distance is set to 1mm, and the ground plane is 0.5mm away. Figure 15: Mesh density setting for the group plane polygon (top left), mesh mirroring on the ground plane (bottom left), normal density mesh (top right), and low density mesh (bottom right) The lower left picture shows the result. The advantage for doing this is that the return current on the ground plane is known to flow under the line, therefore, the mesh more accurately models the distribution of the return current. However, on a board, this setting can cause problems if there are multiple internal planes. The mesh could be mirrored through all layers, causing many poorly formed facets, which is why mesh mirroring is normally either turned off in boards and modules or the mirroring distance is set to 0mm, which effectively turns mirroring off. Mirroring is important for MMIC capacitors, where the mesh on the two conductors of the capacitor should be the same. This necessity for mirroring the mesh of a capacitor is another example of years of experience of mesh development heuristics.

Articles in this issue

Links on this page

view archives of AWR eBooks - Best Practices for Efficient and Effective Planar EM Simulation