EM Options
The EM Options pane controls the behavior of the EM solver. Select Circuit > Settings > [EM Options] to open the EM Options pane. The contents of this pane is different for geometry projects and netlist projects. This topic explains geometry project settings first, and netlist settings last.
Geometry Projects
Speed/Memory
This slider bar allows you to control the memory usage for an analysis by controlling the subsectioning of your circuit. For a complete discussion of this control, please refer to Speed/Memory Control.
Compute Currents
This option outputs current density data which can be viewed using the Current Density Viewer. Please be aware that calculating current density data can increase the time needed for your analysis. For an Adaptive (ABS) Sweep, current density data is only calculated for discrete data points. This option is not available for a DC frequency sweep.
Advanced Section
Clicking on > Advanced expands a section of the dialog box that allows you to control advanced EM options. The advanced section contains three tabs. Each tab is described in detail below.
Analysis Options Tab
De-embed: Enables de-embedding for your project. When this option is enabled, the EM solver de-embeds all ports and produces de-embedded results. For more details about de-embedding, please see De-embedding Overview. This option is enabled by default.
Box resonance info: Outputs box resonance information during an analysis. If a potential box resonance is detected, you will receive a warning message which includes the frequency and type of mode detected. See Box Resonances for more details.
Precision: This checkbox enables a drop-down list which controls the precision of the EM solver calculations. The following is an explanation of each option:
- Double: Double precision (64-bit) calculations are used by the EM solver. This is the default setting and should be used for most cases. On rare occasions, such as a small size or analyzing at low frequencies, numerical precision errors may be encountered.
- Quad: The EM solver will use quad precision (128-bit) for the matrix fill calculations and double precision (64-bit) for the matrix solve calculations. This will result in a slower analysis, but will have minimal impact on the memory requirements. This option should only be used if a double precision simulation results in numerical precision errors.
- Single: The EM solver will use double precision (64-bit) for the matrix fill calculations, and single precision (32-bit) for the matrix solve calculations. This will result in a slower simulation, but uses less memory. Single precision calculations can introduce numerical precision errors at low to moderate frequencies. It should only be used when your computer does not have enough RAM to perform the simulation.
Process derived layers: When this checkbox is selected, Derived Tech Layers are processed by the EM solver. If not selected, Derived Layers are ignored by the EM solver. See Derived Tech Layers for a detailed description of Derived Tech Layers.
Other Options: This entry should only be used at the direction of Sonnet Support personnel.
Configure ABS Tab
The Configure ABS tab contains options that use and settings that affect Adaptive Band Synthesis (ABS).
Options
The options section allows you to enable options that use Adaptive Band Synthesis (ABS). An Adaptive (ABS) Sweep is required to use these options.
Q-Factor accuracy: See Q-Factor Accuracy.
Enhanced resonance detection: See Enhanced Resonance Detection.
ABS Caching Level
The ABS Caching Level determines how much ABS caching data is stored in your project. The default setting (Multi-Sweep plus Stop/Restart) is recommended for most situations. See ABS Caching Level for a description of each choice.
Advanced Subsectioning Tab
The Advanced Subsectioning tab allows you to control the subsectioning of your circuit. Since subsectioning has a direct impact on processing time and accuracy, these settings allow the experienced user a measure of control over the trade-off between them. For detailed information about each option, see Project Settings.
Beowulf Tab
The Beowulf tab contains settings for the Beowulf solver [4].
Matrix Solver: This drop-down list allows you to choose between the Direct Solver or the Beowulf Solver. The default setting is Auto. When set to Auto, Sonnet will choose the solver based on the properties of your circuit.
Preconditioner: This setting only applies if the Beowulf solver is used. Preconditioners are a way to make the Beowulf Solver converge faster. See Preconditioners for details.
Precision: The Precision section only applies if the Beowulf solver is used. See Precision for details.
A Beowulf license is required to use the Beowulf solver.
Estimate Memory Button
Clicking the Estimate Memory button opens the Estimate Memory dialog box which displays an estimate of the memory required to analyze the project and allows you to view the subsectioning. See Viewing Subsections for details.
Netlist Projects
The following EM options are available for netlist projects.
Push frequency sweeps down to all subprojects: Select this checkbox to impose the frequency sweeps used in this netlist on all project elements in the netlist. In other words, the netlist's frequency sweeps are used when analyzing a subproject; the frequency sweeps in any subprojects are ignored. If this option is unchecked, then the subproject uses the frequency sweeps defined in the subproject.
Enable Enhanced Resonance Detection: See Enhanced Resonance Detection for an explanation of this feature. When enabled, the frequency sweeps in any subprojects are ignored.
Other Options: This entry should only be used at the direction of Sonnet Support personnel.