Welcome to CST !

CST STUDIO SUITE

Training Class

Core Module

1
 About CST
Founded in 1992
170 employees
World-wide distribution network
Focus on 3D EM simulation

2
 CST Worldwide

CST West Coast CST of America CST Europe CST China CST of Korea AET Japan

3
 CST Products
Common Easy-To-Use Pre-
CST STUDIO SUITE and Post-processing Engine

RF Simulations
CST CABLE STUDIO

Applications
CST MICROWAVE STUDIO

for Special
CST PCB STUDIO
Our Flagship Product
for RF Simulations CST MICROSTRIPES

CST DESIGN STUDIO

Circuit Simulator
Allows Coupling of 3D Models

CST PARTICLE STUDIO CST MPHYSICS STUDIO CST EM STUDIO

Interaction of EM Fields with Thermal and Mechanical Simulations of Static or

Free Moving Charges Effects of EM Fields Low-Frequency Fields

4
 Built-In Help
Mechanisms

5
 Documentation
<CST_INSTALLATION_DIR>\Documentation\

The introductory books are a good starting

point to learn the workflow of the CST
STUDIO SUITE products.
All books are available as pdf documents
in the "Documentation" subfolder of your
CST installation.

6
 Tutorials
Step-by-Step tutorials are available for CST MICROWAVE STUDIO
and CST EM STUDIO.

7
 Examples Overview
Many pre-calculated examples are available.

Antenna Calculation Examples

8
 Online Help (I)

9
 Online Help (II)
- Links to Online Help -

In almost all dialogs there is a link to the online help documents

which provides you with extensive help for all settings.

Linked page of the online help

Transient solver main dialog
10
 CST Webpage

www.cst.com

11
 CST Support Site
Tutorial Videos

FAQ Section

12
 CST User Forum

Ask your questions. Answers are provided by other users or CST engineers.
13
 CST Customer Support

CST Malaysia
Phone: +60 (3) 7731 5595
Fax: +60 (3) 7722 5595
Email: info@sea.cst.com
Support available from
9am 5pm

14
 CST Training Courses
The training courses for CST STUDIO SUITE provide you with the
knowledge needed for an efficient start with the software.
Currently the following trainings are offered on a regular basis. All
upcoming courses are announced on the CST webpage.

CST STUDIO SUITE CST PARTICLE STUDIO CST MICROSTRIPES

Charged Particle CST MICROSTRIPES
MW & Antenna Training
Dynamics Training Training
2 full days
1 full day 1-2 full day(s)
EMC / SI / PI Training
2 full days CST EM STUDIO
CST CABLE STUDIO
Performance Training LF Applications Training CST PCB STUDIO
1 full day 1 full day Training on Demand

15
 Basic and Advanced
Modeling

16
 Common User Interface
Menu Bar

Tool Bars

Primary
Window

Navigation
Tree

Message

Parameter List Window

17
 Customize Your Environment
E.g., define a shortcut key
to call your favorite macro.

18
 View Options

Rectangle zoom allows to zoom in a rectangular domain.

Change the view by dragging the mouse while pressing the left
button and a key.
ctrl - rotation
shift - in-plane rotation
ctrl+shift - panning
Some other useful options are:
spacebar - reset view to structure,
ctrl+f - reset view,
shift+spacebar - zoom into selected shape,
mouse wheel - dynamic zoom to mouse pointer.

19
 Primitives
Cylinder Cone Torus

Sphere
Rotation

Brick

Elliptical
Hints:
Cylinder Press the tab-key to enter
a point numerically.
Press backspace to delete
a previously picked point.
Extrusion
20
 Picks
Pick a point, an edge, or a face in the structure.

Hints:
Press "s" to activate all pick tools.
Picked Point Picked Edge Picked Face To pick a point by given coordi-
nates, press p and the tab-key.
2nd time picking an element
unselects it.
Pick face
Pick circle Pick face (f)
center (a)
center (c)
Pick corner Clear picked elements (d)
point (p)
Pick edge Edge from
center (m) coordinates
Pick point Pick edge (e)
21 on circle (r)
 Working Coordinate System

The working coordinate system (WCS) allows the use of context

dependent coordinates.
Use to switch on/off the WCS.

Use to rotate the WCS.

Use to move the WCS.

22
 Working Coordinate System

The WCS can be aligned, e.g., with a point, an edge, or a face.

Align the WCS

with a point

Align the WCS

with an edge

Align the WCS

with a face

Press w to align the WCS with the currently selected object.

23
 Working Coordinate System

The position of a WCS can be stored for later use.

24
 Boolean Operations

Boolean operations can be applied to two or more shapes to

create more complex structures.
Sphere

Add Subtract
Brick
Brick + Sphere Brick - Sphere

Intersect Boolean insert

Brick * Sphere Sphere / Brick Brick / Sphere
25
 Curve Modeling Tools Overview (I)

Curves can be used for

structure generation,
thin wire generation,
integration path in post-processing,
healing CAD data.

Basic Curves
Generation

Create new curve

26
 Curve Modeling Tools Overview (II)
Solids can be created from curves.

Creation of a
Sheet from a
Planar Curve

Extrusion of a
Planar Curve

Sweep Curve

27
 Curve Modeling Tools Overview (III)

Solids can be created from curves.

Creation of a
Trace

Creation of
Loft from two
Curves

28
 Rotation of Profile
Rotation Axis

Draw the profile.

Specify rotation angle,
material properties, etc.
Press backspace to delete
the last selected point.
Double click on any corner
point to change its position.

29
 Analytical Modeling (I)
3D curves and faces can be created using analytical expressions.

Enter parameterization

30
 Analytical Modeling (II)
3D curves and faces can be created using analytical expressions.

31
 Loft Operation
Two picked faces can be used to create a new shape by a loft
operation.

Pick two faces.

Choose the properties

of the loft operation.

Preview

32
 Bending
It is possible to bend a sheet on a solid object.

Example:
Creation of a Helix

Sheet

Solid

The solid and the sheet must touch each other.

33
 Blend and Chamfer Edges

Specify angle and width.

Select edges.

Specify radius.

34
 Shell Operation
A solid object can be shelled.
Example:
A waveguide bend consisting of three shapes is shelled.

solid2

solid3
solid1
Create a single shape
by a Boolean add.

Picked faces will be open

after the operation.

35
 Transform Operation
Existing objects can be translated, rotated, mirrored, and scaled.

Translate Scale Rotate

Use the mouse to translate, rotate, or scale objects interactively.

Perform several transformations to the same shape using the Apply
button.
Selecting more than one solid will turn the shape center into the
common center.

36
 Local Modifications Face Modifications

Offset Face: Interactively Move Face: Interactively

move the face of a solid in move the face of a solid in
its normal direction. a coordinate direction.

Local Modifications are especially helpful

when you are working with an imported
CAD model for which the model history is
not available. The "Local Modification"
tools help you to modify such geometries.

37
 Local Modifications Remove Feature

Feature to be removed Pick the feature

Remove the feature

38
 View Options
Several options are available to gain better insight into the structure.

Cutting Plane

Wireframe Mode

39
 View Options
Several options are available to gain better insight into the structure.

Working Plane

Coordinate Axes

40
 Copy / Paste Structure Parts
Ctrl+C stores the selected solids on the active working coordinate
system (WCS) to the clipboard. Ctrl+V pastes the clipboard into the
active working coordinate system.
Copy and paste of structure parts works even between different CST
projects.

Paste the objects in

the new WCS.
Press ctrl+c to copy
objects to clipboard.

Move the WCS.

41
 Align Objects
Copied or imported objects can be aligned with the current model.

Select shape and Select faces to Choose angle. Final Result

choose Align align with.

For copied and imported objects, the alignment is started

automatically.
For shapes selected in the Navigation Tree start by choosing Align
from the Objects menu.

42
 Interactive CAD Modeling Using the Mouse

1. Adjust the Snap width according to the raster of your structure.

2. Use the pick tools, whenever geometrical information is already available.
Pick points to define new shapes / height of extrusion / transform.
Pick edges for rotation axis / to adjust WCS.
Pick face for extrude / rotate / transform / to adjust WCS.
3. Use the local working coordinate system (WCS).
4. Use the keyboard only for new (independent) geometric information
(e.g. points which cannot be picked and do not fit into the snapping raster).

Relative construction via picks and WCS avoids redundant information.

Parameters/Values are entered once and are later referenced via picks.

43
 Notes

44
 Solver Overview
Which solver is best suited to my application?

45
 Which Solver is the Best?
Unique answer to this question is not easily possible as the
performance and accuracy depend on many parameters:
Electrical size and geometry of the problem,
Material models and material parameters used,
Resonant behavior of the model,
Type of the mesh and the boundary conditions,
Architecture of the workstation used for the simulation,
etc.

BUT: Some helpful rules of thumb are available.

The application engineers of CST are available to

discuss the solver choice and the model setup.

46
 Transient Simulation - Behind the Scenes
Excitation Time Signal Output Time Signal

Numerical time integration

of 3D Maxwell equations

Port 1 Port 2

The simulation duration depends on:

1. Duration of input signal (determined by frequency range selected)
2. Duration of output signal (determined mainly by the size and the
resonances of the model under study)
3. Time step width for numerical time integration (determined by the
mesh used to discretize your model)
47
 Frequency Domain Simulation
Behind the Scenes
The steady state behavior of a model is calculated at different
frequency points.

For each frequency

point a linear
equation system
has to be solved.

The intermediate points in broadband results are calculated by an

interpolation.
48
 Time Domain + Frequency Domain

in out in out
Time Domain Frequency Domain Frequency Domain Calculation

in in

out out

TDR S-parameter S-parameter

49
 Solver Choice (I) - Overview
General Purpose Solver (3D-Volume)
Solver Area of Application (Rule of Thumb)
Electrically medium and large sized problems
Transient Broadband
Arbitrary time signals
Narrow band / Single frequency
Frequency
Electrically small to medium sized problems
Domain
Periodic structures with Floquet port modes

Special Solver (3D-Volume): Closed Resonant Structures

Eigenmode Strongly resonant structures, narrow band (e.g. cavities)

Strongly resonant, non radiating structures (e.g. filters)

FD Resonant

Special Solver (3D-Surface): Large Open Metallic Structures

Integral Equation Electrically large structures
(based on MLFMM) Dominated by metal

Asymptotic Solver RCS calculations for electrically very large objects

50
 Solver Choice (II) - Resonances
The following rules of thumb apply:

Weak Resonances Strong Resonances

General Purpose

+AR-Filter
for S-parameter
calculation only

Resonant Fast

F-solver is better suited to strongly resonant applications than T-solver.

51
 Solver Choice (III) - Electrical Size
The following rules of thumb apply:
Structure under study

Electrically Small Electrically Large

With MPI also very large

problems can be solved.

RCS calculations for electrically

very large structures

For electrically very small structures the quasistatic solvers provided in

CST EM STUDIO might be a good choice.

52
 Solver Choice (IV) - Bandwidth
The following rules of thumb apply:

Narrowband Broadband

F-solver and I-solver are better suited to narrowband applications,

while the T-solver
T is better suited to broadband applications.
53
 Specialized Products
In addition to the general purpose solvers of CST MICROWAVE STUDIO
CST offers solvers specialized to certain classes of applications.

CST PCB STUDIO

Specialized solvers for the
simulation of PCB boards.
CST CABLE STUDIO

Specialized solvers for the

simulation of complete cable
harnesses for all kind of EMC
investigations.
CST MICROSTRIPES
Efficient solvers based on the
Transmission Line Matrix (TLM)
method. Contains special
algorithms for EMC analysis.

54
 Notes

55
 Optional Workflow Example
Patch Antenna Array

Purpose 1: Design a single patch using

a parameter sweep & optimization.

Purpose 2: Create a dual patch array

using
a farfield array combination
3D array creation
a beam-forming feeding network
56
 Single Patch

57
 Single Patch Design
Frequency range: 3 8 GHz
Port size:
2*width in y-direction Copper
Substrate (Rogers RT 5880)
5*height in z-direction

40mm
20mm

h = 0.787mm
0.035mm
0.5mm

7.5mm
20mm
w = 2.38mm
40mm
Copper groundplane,
thickness = 0.035 mm

58
 Construction (i)
Choose template: Load materials:

59
 Construction (ii)
Construct the substrate:

Load substrate material

60
 Construction (iii)
Construct the patch:

61
 Construction (iv)

Align WCS with picked

point
Select edge centre

62
 Construction (v)
Construct the feed line

Press Shift-Tab Select edge centre

63
 Construction (vi)
Pick point

Align WCS with picked point

64
 Construction (vii)
Construct the feed
gaps

65
 Construction (viii)
Pick two points to form a
translation vector
Select solid1 by double-
clicking it

66
 Construction (ix)
Transform solid1 to make a copy

67
 Construction (x)
Select component patch
Select component solid1

Hit ENTER to substract

solid1 from patch

68
 Construction (xi)

Pick bottom face of

substrate

Extrude face to make

ground plane

69
 Construct Port

Pick face of feed line

70
 Construct Port
Construct waveguide port

71
 Simulation Settings

Set freq. range Exploit symmetry plane

72
 Simulation
Define monitors (E-, H-, Farfield @ 5.25 GHz)
Start transient solver

73
 Visualize Results

Farfield result

E-field result
74
 Notes

75
 Parameter Sweep

76
 Parameter Sweep Results: S11

77
 Optimization of Single Patch
Optimizer Parameters

Optimizer Goal

78
 Optimizer Results (i)

79
 Optimizer Results (iii)

80
 Farfield Efficiency

Before optimization:
81
 Patch Array

82
 Combine Farfields (1)

83
 Phaseshift = -45 (1R)

84
 Phaseshift = 135 (2L)

85
 Combine Farfields (2)

Transform component1 to make

a copy

Combine ground and

substrate components

86
 Combine Farfields (2)

Construct second port and run transient simulation

without symmetry.

87
 Combine Farfields (2)

1 2

88
 Farfield Results (L)

89
 Farfield Results (R)

90
 Feeding Network Design (DS)
lg/4

lg/4
Z0

Z0/sqrt(2)

91
 DS MWS co-simulation

3D MWS model fed with

DS circuit network

92
 Notes

93
 Definition of Ports

94
 Available Port Types
Ports for S-Parameter Computation

Discrete Ports Waveguide Ports

(Lumped Element) (2D Eigenmode Solver)

Input: Knowledge of TEM Mode and Input: Area for eigenmode solution
line impedance is required. Output: Pattern of E- and H-field,
Output: Voltage and current line impedance,
Propagation constant
Discrete ports can be used for TEM-like modes, not for higher order
modes (cutoff frequency > 0).
Waveguide ports provide a better match to the mode pattern as well
95
as higher accuracy for the S-parameters.
S
 Discrete Ports
S-Parameter Port Current Port Voltage Port

Voltage or current source with

internal resistance

Coaxial Microstrip Stripline Coplanar waveguide

96
 Discrete Edge Port Definition

Pick two points, or pick one point and a face,

or enter coordinates directly (not recommended).

Select port type

and impedance.

97
 Discrete Face Port Definition

Pick two edges or one edge and a face.

Select port type

and impedance.

98
 Available Port Types
Ports for S-Parameter Computation

Discrete Ports Waveguide Ports

(Lumped Element) (2D Eigenmode Solver)

Input: Knowledge of TEM Mode and Input: Area for eigenmode solution
line impedance is required. Output: Pattern of E- and H-field,
Output: Voltage and current line impedance,
propagation constant
Discrete ports can be used for TEM-like modes, not for higher order
modes (cutoff frequency > 0).
Waveguide ports provide a better match to the mode pattern as well
99
as higher accuracy for the S-parameters.
S
 Port Definition (I) Closed Structures
Typically, waveguide ports are defined based on a geometric object. Use the
pick tools to select a unique port plane.

The port size is equal to the smallest rectangular area which includes all picked objects.

100
 Port Definition (II) Open Structures

1. Pick three points.

2. Enter port menu .
3. Adjust additional
port space.

101
 Port Definition (III) - Backing
For the I-solver and the F-solver waveguide ports must be backed with
a PEC solid (or by electric boundaries).
Pick port using
the pick tools.

Port backed with PEC solid.

102 Extrude the port plane.

 Notes

103
 Materials
&
Boundary Conditions

104
 Basic Materials
Define a new material or load materials from the large material database.

Material Types
PEC = Perfect Electrical Conductor ( )
Normal: General material model. This is
typically used for dielectric materials.

Anisotropic: Permittivity and permeability

depend upon the spatial direction.
Lossy Metal: Model for conductors with .
Corrugated Wall: Surface impedance model.

Ohmic Sheet: Surface impedance model.

105
 Material Database

Loaded materials are available

for the creation of new shapes.

106
 Lossy Metal
Why is it required?
Sampling of skin depth would require very fine mesh steps at
the metal surface when defining conductor as a normal material
(skin depth for copper at 1 GHz approx. 2 m).
This results in a very small time step, which leads to a very long
simulation time.

Solution:
1D model which takes skin depth into account without spatial
sampling.

107
 Boundaries
CST MWS uses a rectangular grid system, therefore, also the complete calculation
domain is of rectangular shape 6 boundary surfaces have to be defined at the
minimum and maximum position in each coordinate direction (xmin, xmax, ymin,
ymax, zmin, zmax).

Example: T-Splitter

xmin
ymax

zmin

zmax

ymin

xmax
108
 Boundary Settings (I)
Seven different settings are available.

109
 Boundary Settings (II)
Electric Boundaries (default setting): No tangential electric field at surface.

Magnetic Boundaries: No tangential magnetic field at surface. Default

setting for waveguide port boundaries.

Open Boundaries: Operates like free space Waves can pass this boundary
with minimal reflections. Perfectly matched layer (PML) condition.

Open (add space) Boundaries: Same as open, but adds some extra space for
far field calculation (automatically adapted to center frequency of desired
bandwidth). This option is recommended for antenna problems.

Conducting Wall: Electric conducting wall with finite conductivity (defined

in Siemens/meter).

110
 Boundary Settings (III)
Periodic Boundaries: Connects two opposite boundaries where the calculation
domain is simulated to be periodically expanded in the corresponding direction.
Thus, it is necessary that facing boundaries are defined as periodic.
The resulting structure represents an infinitely expanded antenna pattern,
phased array antennas. F! (hexahedral mesh), T! + 0 phase shift

Unit Cell: Used with F! solver, tetrahedral mesh, similar to F! periodic

boundary with hexahedral mesh. A two dimensional periodicity other than
in direction of the coordinate axes can be defined. If there are open
boundaries perpendicular to the unit cell boundaries, they are realized by
Floquet modes, similar to modes of a waveguide port .

111
 Boundaries: Symmetry Planes
Three different settings are available.
Three possible symmetry planes.

112
 Notes

113
 Meshing Basics

114
 How to Get a Proper Mesh?
Question: How does a proper mesh look like and what are the
best settings to get it?

Answer: This depends on your problem under study as well as

the type of result you are interested in.

However, there are some rules of thumb:

For several classes of application (e.g. antennas, PCB boards
etc.) there are some common properties a "good" mesh
possesses (project templates make use of this fact).
It is known that the results become more accurate when the
mesh is refined (automatic mesh refinement is based on this
knowledge).
Geometry and material of the model influences the behavior of
the EM fields (fixpoints, material based meshing, and other
special techniques are based on this knowledge).

115
 Hierarchy of Mesh Settings

Global Mesh Properties General settings usually done by project

template. Global settings for mesh controls of
automatic meshing algorithms.

Local Mesh Properties Special settings (fine-tuning) to adjust the

global mesh better to the model under study.
Defined per shape or per material.

Local mesh properties have precedence over global mesh properties.

116
 Mesh Generation - A Typical Workflow
Select Project Template This adjusts the global mesh properties to
values which we found to be a good starting
point for a certain area of application.

Global Mesh Settings Optimize the global mesh settings for the
geometry of your model.

Local Mesh Settings Fine tune the mesh (if necessary) to meet the
really specific requirements of your model.

Perform Simulation Start the solver and perform a convergence

study (e.g. using adaptive mesh refinement).

Simulations and mesh studies provide insight

Results about the dependency of the results on the
117 mesh settings.
 Project Templates
A project template makes some basic settings for a new project. A
project template can be applied to an already existing project.

Information about the

settings the template
will apply.

Template Title
(Area of Application) Initial Mesh Settings

118
 Automatic Mesh Refinement (I)
It is known that the numerical solution calculated by the solvers converges to
the analytical solution if the grid is sufficiently refined.
The automatic mesh refinement in CST tries to refine the initial mesh in a
clever way such that the results are accurate.

119
 Automatic Mesh Refinement (II)

The results for different meshes during an adaptive mesh

refinement are shown in the "Navigation Tree".

120
 Hexahedral Meshing for
Transient Simulations

121
 Hexahedral Meshing - Overview

1. Hexahedral Mesh Configuration Options

2. Some Meshing Guidelines

2.1 Some Representative Meshes for Common Structures

2.2 Meshing Pitfalls

3. Influence of the Mesh on Simulation Performance

122
 Hexahedral Mesh (I) - Mesh View
Mesh lines in one
mesh plane are shown
in the 3D view.
View mesh.
Mesh controls are
displayed in the mesh
view.

Information about mesh plane.

The total number of mesh Corner Fixpoints

cells is displayed in status bar. Correction
123
 Hexahedral Mesh (II) - Global Settings

Absolute and frequency

dependent setting to
determine the largest
mesh step.

Automatically create
and use mesh controls.
Settings to limit the
Strongly recommended!
size of the smallest
mesh step.

124
 Hexahedral Mesh (III) - Global Settings
Largest Mesh Step - "Lines per Wavelength"

"Lines per wavelength" is based on the

upper limit of the frequency range.
Thus, increasing the upper frequency limit
usually leads to a finer mesh.
125
 Hexahedral Mesh (IV) - Global Settings
Largest Mesh Step - "Lower Mesh Limit"

"Lower Mesh Limit" is based on the

dimensions of the computational domain.
The diagonal of the smallest boundary
face of the comp. domain is divided by
this number. Result is used as the max.
126 mesh step width allowed in the model
 Hexahedral Mesh (V) - Global Settings
Smallest Mesh Step - "Mesh Line Ratio Limit"
The time needed to complete a time domain simulation heavily depends on the size
of the smallest mesh step (see later in section "Performance Aspects of Meshing").

The size of the minimum mesh step can be

limited using the "Mesh Line Ratio Limit" or the
"Smallest Mesh Step" setting.

Mesh lines are

inserted at Mesh Line Ratio Limit
fixpoints
fixpoints.

The "Mesh Line Ratio Limit" specifies the

maximum value allowed for the ratio of the
maximum mesh step width to the minimum
mesh step width.
127
 Hexahedral Mesh (V) - Global Settings
Smallest Mesh Step - "Smallest Mesh Step"
The time needed to complete a time domain simulation heavily depends on the size
of the smallest mesh step (see later in section "Performance Aspects of Meshing").

Smallest Mesh Step

The "Smallest Mesh Step" specifies the minimum

value allowed for the minimum mesh step
width in terms of the units defined in your
project.
Note: If the settings for "Steps per Wavelength"
or "Lower Mesh Limit" lead to a smaller
then the "Smallest Mesh Step" setting is
ignored.
128
 Hexahedral Meshing - Overview

1. Hexahedral Mesh Configuration Options

2. Some Meshing Guidelines

2.1 Some Representative Meshes for Common Structures

2.2 Meshing Pitfalls

3. Influence of the Mesh on Simulation Performance

129
Representative Meshes (I) - Minimal Requirements

The gap between inner and outer conductor

should be resolved by at least one mesh cell.
Partially filled cells are handled with PBA/FPBA
technique.

Coaxial Line
Depending on the thickness and the
1-2 mesh lines permittivity of the substrate the number
of mesh lines should be at least as shown
in the picture.
2-3 mesh lines It is NOT necessary to resolve the
(depends on thickness) thickness of the microstrip line by the
mesh.

Microstrip Line
130
Representative Meshes (II) - Minimal Requirements

The gap between multiple strip lines should be

resolved by at least one or two mesh cells.

Parallel Microstrip Lines

A discrete port must be discretized by at least

one mesh cell.

Discrete Ports
131
 Meshing Pitfalls - Staircase Cells (I)
Cells which contain more than two metallic
material boundaries are completely filled
with PEC (staircase cells).

A warning is shown by the

solver to inform you of this
modification.

Staircase cells are shown in the

mesh view.

132
 Meshing Pitfalls - Staircase Cells (II)

Staircase cells must be avoided if

they influence the electrical
behavior of the model, i.e. if they
introduce shortcuts.
Example: Shortcut between two
microstrip lines is introduced by a
staircase cell.

Staircase cells which do not change

the electrical behavior of a model
are usually OK.

Example: Staircase cell at

a wire in free space.
133
 Online Help - PBA and TST

PBA TST
Whenever a mesh cell cuts more than two metallic material
boundaries the cell is filled with PEC material (staircase cell).
Quite often such cells do not influence the simulation result
much, but if they introduce shortcuts (as shown on the previous
slide) this might be critical.
134
 Hexahedral Meshing - Overview

1. Hexahedral Mesh Configuration Options

2. Some Meshing Guidelines

2.1 Some Representative Meshes for Common Structures

2.1 Meshing Pitfalls

3. Influence of the Mesh on Simulation Performance

135
 Hexahedral Meshing Performance (I)
For stability, the time step of the numerical quadrature is determined by the
smallest mesh step. Increasing the smallest mesh step will increase the
time step.

Smallest Mesh Step

tiny t: slow big t: fast

t t

The smaller the smallest mesh step width, the smaller the time
step for the numerical time integration.
136
 Hexahedral Meshing Performance (II)
The smallest mesh step in a model can be visualized in the mesh
view.

137
 Hexahedral Meshing Guidelines - Summary

Select a proper project template for your application to get good

initial mesh settings.

Perform an adaptive mesh refinement to find a good mesh.

Fine tune the mesh if necessary using the local mesh settings.

Try to avoid critical cells. Quite often they are an indicator that the
mesh is too coarse at least in some regions.
Try to avoid to use a mesh with a very high mesh line ratio limit.
Consider using subgrids for models which require a very fine mesh at
localized positions.

138
 Transient Simulation - Memory Consumption
- Memory-Consumption versus Mesh Size -

Some rules of thumb are:

A structure with open boundaries and material losses requires
about 1 GB RAM to handle 3-4 million mesh cells.
A structure with closed boundaries and without material losses
requires about 1 GB RAM to handle 5 million mesh cells.
Subgridding:
The subgridding feature starts to be efficient when the mesh
cell reduction factor is larger than 3.
(Macros Calculate Subgridding Meshcell Factor)

139
 Tetrahedral and Surface Meshing
for Frequency Domain Simulations

140
 Global Mesh Properties
Steps per wavelength: This value refers to the
highest frequency of the simulation. It defines the
minimum number of mesh cells that are used for a
distance equal to this wavelength.

Minimum number of steps: This value controls the

global relative mesh size and defines a lower bound
for the number of mesh cells independently of the
wavelength It specifies the minimum number of
wavelength.
mesh edges to be used for the diagonal of the model
bounding box.

Note: A tetrahedral mesh requires a valid ACIS model.

(HEX mesh even works with INVALID ACIS model...)
141
 Tetrahedral / Surface Mesh (I)
-Global Mesh Settings -

"Steps per wavelength" is based on the

upper limit of the frequency range.
Thus, increasing the upper frequency limit
usually leads to a finer mesh.
142
 Tetrahedral / Surface Mesh (II)
-Global Mesh Settings -

"Min. number of steps" allows to refine

the mesh globally independently of the
frequency range settings.
It specifies the minimum number of
mesh edges to be used for the diagonal
143 of the model bounding box.
 Mesh Generation Method
The method for surface and volume meshing can be chosen.
Geometry accuracy: If the defined
or imported geometry is less
accurate than the default tolerance
1e
1e-6, it is recommended to select a
larger tolerance. Otherwise artificial
shapes might arise or the model
preparation might fail.
144
General purpose: A simple surface mesh
generation which is adequate in most cases.
Fast (for complex structures): Especially suited
to meshing large or complex structures. If used
together with (tetrahedral) volume mesh
generation, this method can be combined only
with Delaunay volume mesh generation.
Delaunay: Fast tetrahedral volume meshing
method (recommended).
Advancing Front: An alternative method to
generate a volume mesh. Advantageous in some
cases (like thin layers), because the surface
mesh can be generated more flexible than with
Delaunay, that is, it can be altered during the
mesh generation if necessary. This method is
available only in combination with the general
purpose surface mesh generation.
 Curvature Refinement (I)

30 100

default = 100
If cylinders are
still not well
discretized,
increase it
to, e.g., 200-300.

145
 Curvature Refinement (II)

The Curvature refinement ratio specifies the ratio of

the maximum deviation (d) of the surface mesh from the
actual shape of the structure divided by the edge length
(h) of the surface triangle (as shown in the picture above)..
Smaller values lead to better approximation of curved
objects
objects.

Volume optimization:
optimization If this field is checked
(recommended), the mesh connectivity of the
preliminary volume mesh is changed to improve the
mesh quality.

Volume smoothing:
smoothing If this field is checked
(recommended), the position of mesh vertices will
be changed in order to enhance the mesh quality.

146
 Adaptive Mesh Refinement
Multi-frequency adaptive mesh refinement
The adaptation frequency samples are sequentially processed before
the broadband sweep.
Example: Diplexer

Mesh adaptation at 75.1 GHz and 77 GHz.

Initial mesh Optimized mesh

147
 Open Discussion

148
 Notes

149