Slides For Half Day Momentum PDF

Agilent EEsof EDA
Presentation on ADS Momentum
This document is owned by Agilent Technologies, but is no longer kept current and may contain obsolete or
inaccurate references. We regret any inconvenience this may cause. For the latest information on Agilent’s
line of EEsof electronic design automation (EDA) products and services, please go to:
www.agilent.com/find/eesof
ADS Momentum
A Half-Day Seminar
Keefe Bohannan
Agilent EEsof Applications Engineer
April 2003
Agenda for Half-day Momentum Seminar
30 minutes Brief overview of Getting Started with Momentum
• Creating/importing artwork in ADS Layout
• Momentum versus Momentum RF
• Creating substrate stack-ups and mapping layout layers as metallization layers
• Placing and defining ports
• Defining mesh parameters
30 minutes Overview of Viewing and Using Momentum Results

• Momentum Datasets
• Momentum Visualization: currents, fields, s-parameters, gamma, Z0
• ADS Data Display: s-, y-, and z-parameters, reactance (L/C), Q, etc.
60 minutes Advanced Topics [Part 1]

• Momentum Co-simulation (EM/circuit co-simulation) using Layout Components
• Momentum Co-optimization (EM/circuit co-optimization) using Layout Components
• Thick conductor simulations {LAB}
• Spiral Inductor simulations {LAB}
15 minutes Break

• Advanced Model Composer (AMC)
• Advanced Model Composer (AMC) {LAB}
15 minutes Final Q&A Session
4 hours 15 minutes
Momentum Seminar momentum_01_01 Page 2

What is meant by Planar EM simulation ?
• Substrate - multiple dielectrics

• Metals - traces on different layers forming component and/or thin film
interconnect
• Vias - connecting different layers
• Method of Moments technique
• Sometimes referred to as 2.5D
• It does NOT include:

• Arbitrary 3D structures
• Horn Antennas

Why are Planar EM Simulators used ?
• No simple analytical model exists

• Coupling between conductors or layers is significant
• Arbitrary planar geometry
• Narrow frequency response not captured by analytical models
• Radiation patterns of planar antennas
• CPW transmission lines
• When full 3D analysis would take too long

How are Planar EM Simulators used ?
• Layout driven
• Created entirely within layout,
• Schematic-to-Layout translation, OR
• Import – (DXF, GDSII, etc.)
• Momentum interface within ADS Layout
• Mode > Substrate/Metallization > Port > Mesh >
Simulation > Component > Optimization
• Outputs
• S-parameters
• Current visualization

Created entirely within layout
Creating/importing artwork in Layout Schematic-to-Layout translation
Import – (DXF, GDSII, etc.)

• Everything is placed on a layer
• There are 39 default layers
• User may add new layers, remove existing layers, or
modify layer names and properties
• User may define name, color, pattern, shape display
(outlined/filled), and line style associated with layers
• Layers may be set to be visible/invisible,
selectable/unselectable, and insertable/uninsertable
• Items on unselected layers may not be selected /
edited / moved / deleted
• Only one layer is “insertable” at a time
• Objects can only be created, copied, or moved
TO the “insertable” (they can be copied or
moved FROM any selectable layer)

• Preferences which apply to things not yet placed: • Preferences for schematic and layout window are set
• Trace, Placement, Entry/Edit, Units/Scale, separately
Component Text, Text • Preferences Setting are saved to file in project
• Preferences which apply to things already placed: • layout.prf and schematic.prf
• Select, Grid/Snap, Pin/Tee, Display, Layout • Preferences files from other projects may be read in
units Options > Preferences Read... button





Simplify hierarchical designs by

using “Generating Artwork”
(flattens hierarchy, but retains
multi-layer layout)

How are Planar EM Simulators used ?
• Layout driven
• Created entirely within layout,
• Schematic-to-Layout translation, OR
• Import – (DXF, GDSII, etc.)
• Momentum interface within ADS Layout
• Mode > Substrate/Metallization > Port > Mesh >
Simulation > Component > Optimization
• Outputs
• S-parameters
• Current visualization

For more information about Agilent Email Updates www.agilent.com
Agilent EEsof EDA, visit:
www.agilent.com/find/emailupdates For more information on Agilent Technologies’
products, applications or services, please
Get the latest information on the
www.agilent.com/find/eesof products and applications you select.
contact your local Agilent office. The
complete list is available at:
www.agilent.com/find/contactus
Agilent Direct
Americas
www.agilent.com/find/agilentdirect Canada (877) 894-4414
Quickly choose and use your test Latin America 305 269 7500
equipment solutions with confidence. United States (800) 829-4444
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008
Europe & Middle East

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
Switzerland 0800 80 53 53
United Kingdom 44 (0) 118 9276201
Other European Countries:
Revised: March 27, 2008
Product specifications and descriptions

in this document subject to change
without notice.
© Agilent Technologies, Inc. 2008

Printed in USA, May 19, 2003
5989-9596EN
Agilent EEsof EDA
Detailed Presentation on Momentum - Part 1 of 3
Solution process
Using Momentum • Select Mode
• Substrate definition
• Port Setup
• Mesh Generation
• Planar Solve
• Enable regular Momentum or Momentum RF • Display Results
• Define Substrate and Metallization (pre-compute option)

• Modify the type and impedance of ports
• Describe a possible Substrate enclosure
• Create/modify Momentum Component to be used in EM/circuit co-simulation
or co-optimization
• Define Mesh parameters (pre-compute option)
•Setup and Perform a Momentum simulation (planar solve)
• Setup and Perform a Momentum optimization (geometric perturbation based)
• Display Visualization (S-parameters, current density, transmission line
parameters) and Radiation patterns
• Export 3D files for HFSS

Using Momentum: Selecting the Analysis Mode
Click this submenu to toggle the analysis mode
• Solution process • Momentum MomentumRF

• MomentumRF Momentum
• Select Mode
• Port Setup
• Mesh Generation
• Planar Solve
• Display Results

Momentum versus MomentumRF: A Snapshot
Momentum features: Momentum RF features:
• Full-Wave EM Simulation • Quasi-Static EM Simulation
• Rooftop Basis Function • Star/Loop Basis Functions
• Rectangular and Triangular Cells • Polygonal cells
• For most passive geometry • Best for geometrically complex designs
• Full accuracy for all circuit sizes • For electrically small designs (≤ λ/2)
• No inherent upper frequency limit • Upper frequency depends on size
• Potential instability at f <kHz to MHz • Results stable down to DC
• Port Calibration • Port Calibration
• Box and Waveguide inclusion • No Box / Waveguide Modes
• Includes all radiation modes • For designs that don’t radiate
• Display 2D and 3D Radiation Patterns • No Radiation Patterns
• Great for 1st pass results, even for large designs (> λ/2)
• Simulation time and memory decrease by ~10X-25X

Momentum versus MomentumRF
Electrically Small condition for Momentum RF…
How do I know?
Status window provides rule of thumb frequency
for which the structure is “electrically small”
D ≤ λ/2

Planar EM Simulation Basics
Physical Design
• Substrate
• Metallization
• Ports
B1(r) B2(r) B3(r)

Method of Moments
• Meshing I1 I2 I3
• Rooftop functions
λ/10
Hey…where did J(r) = I1B1(r) + I2B2(r) + I3B3(r)
this equation
come from?

Maxwell’s Equations
∇×E = -･B/･t Faraday’s Law
∇×H = J + ･D/･t Ampere’s Law
∇·D = ρ Gauss’s Law
∇·B = 0 No Name (Gauss’s Law for Magnetism)
where
E = Electric Field Intensity Vector
H = Magnetic Field Intensity Vector James C. Maxwell
D = Electric Flux Density (Electric Displacement Vector)
B = Magnetic Flux Density Vector
If one then transforms these equations to the integral form, the mixed potential integral equation in very general
form as a linear integral operator equation follows:
Here, J(r) represents the unknown surface currents and E(r) the known excitation of the problem. The Green's dyadic of the
layered medium acts as the integral kernel. The unknown surface currents are discretized by meshing the planar
metallization patterns and applying an expansion in a finite number of subsectional basis functions B1(r), ..., BN(r):
Ohhhh…sorry I
asked.
Planar EM Simulation Basics
Method of Moments B1(r) B2(r) B3(r)
Maxwell’s Equations I1 I2 I3
Matrix Equation ≤λ/10

[Z].[I]=[V] L13
I1 L12 I2 L23 I3
Equivalent Circuit L11 R22 L22 L33

C11 C22
[Z] = [R] + jω[L] + 1/jω [C]-1
C12

Fullwave versus Quasi-Static: Fullwave
1 e − jkR
Fullwave EM R
•Fullwave electric & magnetic Green’s functions
Matrix Equation •Includes space and surface radiation

[Z].[I]=[V] •[L(w)] & [C(w)] are complex and frequency dependent
•[Z(w)] matrix reload CPU intensive
Equivalent Circuit
[Z] = [R] + jω[L(ω)] + 1/jω [C(ω)]-1
[S]

Fullwave versus Quasi-Static: Quasi-Static
1 e − jkR ≈ R1 (1 − jkR + ...)

Quasi-Static EM R
•Electro- and magneto-static Green’s functions
Matrix Equation •Near field / low freq approximation
[Zo].[I]=[V] L(w) = L0 + L1wR + L2(wR)2 + …
C(w) = C0 + C1wR + C2(wR)2 + …
Equivalent Circuit • Neglects far field radiation
[Zo] = [R] + jω[Lo] + 1/jω [Co]-1 • [L0] & [C0] are real and frequency independent
• [Z0] matrix reload very fast
[S]

A Summary of Effects Included
Layout
DC Spice Momentum RF Momentum MW
Spice model S parameters S parameters
RF MW
• quasi-static inductance . . . . . .
• quasi-static capacitance . . . . .
• DC conductor loss (s) . . . . . . . .
• DC substrate loss (s) . . . . . . . .
• dielectric loss (tan d) . . . . . . . . . . . . . . . . . . . . . . . . . . .
• skin effect loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
• substrate wave radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
• space wave radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Using Momentum: Creating Substrate Stack-ups and
Mapping Layout Layers as Metallization Layers
• Solution process
• Select Mode
• Port Setup
• Mesh Generation
• Planar Solve
• Display Results

Once you have created or imported your artwork…

…be sure to define (or open) your substrate stack-up and map the metallization layers

Greens Function Substrate Calculation Time
Student’s Guide A-36

A note on layout layer conductivity
Conductivity defined as:

• Perfect Conductor (lossless)
• σ (Real, Imaginary)
• σ (Real, thickness)
• Impedance (Real, Imaginary)
The parameters selected are applied
toward a conductor loss algorithm, this
does NOT affect the layout thickness

Using Momentum: Creating Substrate Stack-ups and Mapping
Layout Layers as Metallization Layers: Loss Model used in Strip Conductors
• Momentum treats all conductors as having zero thickness. However, the conductivity and thickness can be specified to approximate
frequency dependent losses in the metallization patterns.
• Momentum uses a complex surface impedance for all metals that is a function of conductor thickness, conductivity, and frequency.
• At low frequencies, current flow will be approximately uniformly distributed across the thickness of the metal. Momentum uses
this minimum resistance and an appropriate internal inductance to form the complex surface impedance.
• At high frequencies, the current flow is dominantly on the outside of the conductor and Momentum uses a complex surface
impedance that closely approximates this skin effect.
• At intermediate frequencies, where metal thickness is between approximately two and ten skin depths, the surface impedance
transitions between those two limiting behaviors.
• This surface impedance is added to the Method of Moments approach that is used for Momentum in general.
• The formula used is a combination of a high-frequency conductivity and a low-frequency bulk resistivity. The formula is such that both
approaches (LF bulk behavior HF surface impedance) transition seamlessly.
• The formula is: We will examine a

• Z = coth(γ) * Zc “thick conductor”
•where Zc = the HF impedance and coth(γ) is the correction for finite thickness
• Zc = 0.5 * sqrt(j * µ0 * ω/(σ + j * ε0 * ω ))
method later in this
• γ = 0.5 * thickness * sqrt(j * µ0 * ω * (σ + j * ε0 * ω)) seminar
•where ω = 2 * π * f
•and σ = conductivity = 1/resistivity [in Siemens/meter]
• The meshing density can affect the simulated behavior of a structure. A more dense mesh allows current flow to be better represented
and can slightly increase the loss. This is because a more uniform distribution of current for a low density mesh corresponds to a lower
resistance
Using Momentum
• Select Mode
• Port Setup
• Mesh Generation
• Planar Solve
• Display Results

Placing and Defining Ports
Considerations
Keep the following points in mind when adding ports to circuits to be simulated using Momentum:
• The components or shapes that ports are connected to must be on layout layers that are mapped to
metallization layers that are defined as strips or slots. Ports cannot be directly connected to vias.
• Make sure that ports on edges are positioned so that the arrow is outside of the object, pointing
inwards, and at a straight angle.
• Make sure that the port and the object you are connecting it to are on the same layout layer. For
convenience, you can set the entry layer to this layer; the Entry Layer listbox is on the Layout tool
bar.
• A port must be applied to an object. If a port is applied in open space so that is not connected to an
object, Momentum will automatically snap the port to the edge of the closest object. This will not be
apparent from the layout, however, because the position of the port will not change.
• If the Layout resolution is changed after adding ports that are snapped to edges, you must delete
the ports and add them again. The resolution change makes it unclear to which edges the ports are
snapped, causing errors in mesh calculations.
Note Do not use the ground port component (Component > Ground) in circuits that will be
simulated using Momentum. Either add ground planes to the substrate or use the ground reference
ports.
(Ground port component toolbar button: )

Description of Momentum Port Types
Port Type General Description Placement Type of layer
• Single Calibrated to remove mismatch at port Edge Strip or

(default) boundary (might also call this a Slot
transmission line port)
•Internal Not calibrated (might also call this a Edge or Strip

direct excitation port) Surface
• Differential Two ports with opposite polarity Edge Strip
• Coplanar Two ports with opposite polarity Edge Slots
• Common Mode Two ports with the same polarity Edge Strip
• Ground Ref. An explicit ground reference for a Edge or Strip

Single or Internal port. Surface
CPW NOTE: For finite ground planes, use Ground Reference ports and Internal port on center conductor.

Single Port Properties
• It is connected to an object that is on either a strip or slot metallization layer.
• It can be applied only to the edge of an object.
• The port is external and calibrated. The port is excited using a calibration process that
removes any undesired reactive effects of the port excitations (mode mismatch) at the port
boundary. This is performed by extending the port boundary with a half-wavelength
calibration (transmission) line. The frequency wavelength selected during the mesh or
simulation process is used to calculate the length of the calibration line. For more information
about the calibration process, refer to "Calibration and De-embedding of the S-parameters"
on page A-7 in the Momentum manual.
• The port boundary can be moved into or away from the geometry by specifying a reference
offset. S-parameters will be calculated as if the port were at this position.
• When two or more single ports are on the same reference plane, coupling effects caused by
parasitics affects the S-parameters. The calibration process groups the ports so that any
coupling in the calibration arms is included in the S-parameter solution.
• If the port is connected to an object on a strip layer, the substrate definition must include at
least one infinite metal layer: a top cover, ground plane, or a slot layer, or a ground
reference must be used in addition to the port.
• If the port is connected to an object that is on a slot layer, the port has polarity.
Tip It is not necessary to open the Port Editor dialog box to assign this port type. Any port
without a port type specified is assumed to be a single port.

Defining a Single Port
• Choose Momentum > Port Editor.
• Select the port that you want to assign this type to.
• In the Port Editor dialog box, under Port Type, select Single.
• Enter the components of the port impedance in the Real and Imaginary fields,
and specify the units.
• You can shift the port boundary, also referred to as the port reference plane.
Shifting the boundary enables a type of de-embedding process that effectively
adds or subtracts electrical length from the circuit, based on the characteristic
impedance and propagation characteristic of the port. Enter the offset in the
Reference Offset field, and select the units. A positive value moves the port
boundary into the circuit, a negative value moves the port boundary away from
the circuit.
• Click Apply to add the definition to the port.

Single Port: Avoiding Overlap (of calibration arm)
Be aware that when using single ports, the calibration arm applied to a port may be
long enough to overlap another element in the circuit. In this case, the port will be
changed to an internal port type, and no calibration will be performed on it. If this
occurs, a message will be displayed during simulation in the Status window indicating
the change.

Single Port: Applying Reference Offsets
Reference offsets enable you to reposition single port types in a layout and thereby adjust electrical lengths in a layout, without
changing the actual drawing. S-parameters are returned as if the ports were placed at the position of the reference offset.
Why Use Reference Offsets?

The need to adjust the position of ports in a layout is analogous to the need to eliminate the effect of probes when measuring
hardware prototypes. When hardware prototypes are measured, probes are connected to the input and output leads of the Device
Under Test (DUT). These probes feed energy to the DUT, and measure the response of the circuit. Unfortunately, the measured
response characterizes the entire setup, that is, the DUT plus the probes. This is an unwanted effect. The final measurements should
reflect the characteristics of the DUT alone. The characteristics of the probes are well known, so measurement labs can
mathematically eliminate the effects of the probes, and present the correct measurements of the DUT.
There are significant resemblances between this hardware measurement process and the way Momentum operates. In the case of
Momentum, the probes are replaced by ports, which, during simulation, will feed energy to the circuit and measure its response. The
Momentum port feeding scheme also has its own, unwanted effect: low-order mode mismatch at the port's boundary, although this is
eliminated by the calibration process. However, in order for this calibration process to work well, it is necessary that the fundamental
mode is characterized accurately. This can only be accomplished when the distance between the port boundary and the first
discontinuity is sufficiently large, that is, there exists a feedline that is long enough to provide this distance.

Single Port: Allowing for Coupling Effects
If you have two or more single ports that lie on the same reference plane, the calibration process will
take into account the coupling caused by parasitics that naturally occurs between these ports. This
yields simulation results that more accurately reflect the behavior of an actual circuit.
The figure below helps illustrate which ports will be grouped in order for the calibration process to
account for coupling among the ports. In this setup, only the first two ports will be grouped, since the
third port is an internal port type and the fourth port is on a different reference plane. Note that even
though the second port has a reference offset assigned to it, for this process they are considered to be
on the same plane and their reference offsets will be made equal.
If you do not want the ports to be grouped, you must add a small thickness of metal to the edge of the
object that one of the ports is connected to. The ports will no longer be on the same plane, and will not
be considered part of the same group.

Internal Port Properties
• Internal ports enable you to apply a port to the surface of an object in your design. By using internal
ports, all of the physical connections in a circuit can be represented, so your simulation can take into
account all of the EM coupling effects that will occur among ports in the circuit. These coupling effects
caused by parasitics are included in your simulation results because internal ports are not calibrated.
• You should avoid geometries that allow coupling between single and internal ports to prevent incorrect S-
parameters.
• An example of where an internal port is useful is to simulate a bond wire on the surface on an object.
Another example of where an internal port is necessary is a circuit that consists of transmission lines that
connect to a device, such as a transistor or a chip capacitor, but this device is not part of the circuit that
you are simulating. An internal port can be placed at the connection point, so even though the device is
not part of the circuit you are simulating, the coupling effects that occur among the ports and around the
device will be included in your simulation.
• Internal ports are often used in conjunction with ground references.

Internal Port Properties
• It can be applied to the interior of a circuit by applying it to the surface of an object.

• It can be applied to the edge of an object.
• It can be applied to objects that are on strip layers only.
• The orientation of the port is not considered if it is on the surface of an object. (For a
description of port orientation, refer to "Adding a Port to a Layout" on page B-5 in
the Momentum manual.)
• No calibration is performed on the port. Because no calibration is performed on the
port, the results will not be as accurate as with a single port. However, the
difference in accuracy is small.
Defining an Internal Port

•Choose Momentum > Port Editor.
•Select the port that you want to assign this type to.
•Click Apply.

Illustration of Internal Port Excitation: Direct Point Feed
direct excitation point feed

Illustration of Internal Port Excitation: Direct Line/Edge Feed
direct excitation
line feed

Differential Port Properties
Differential ports should be used in situations where an electric field is likely to build up
between two ports (odd modes propagate). This can occur when:
• The two ports are close together
• There is no ground plane in the circuit or the ground plane is relatively far away
• One port behaves (to a degree) like a ground to the other port, and polarity between the
ports is developed.
• The ports are connected to objects that are on strip metallization layers.
• The electric field that builds up between the two ports will have an effect on the circuit
that should be taken into account during a simulation. To do this, use differential ports.
Differential ports have the following properties:

• They can be applied to objects on strip layers only.
• They are assigned in pairs, and each pair is assigned a single port number.
• Each of the two ports is excited with the same absolute potential, but with the opposite
polarity. The voltages are opposite (180 degrees out of phase). The currents are equal but
opposite in direction when the ports are on two symmetrical lines, and the current
direction is approximated for other configurations.
• The two ports must be on the same reference plane.

Differential Port Numbering
Note: Port numbers for differential ports are treated in the following manner: on the layout, you will continue
to see the port numbers (instance names) that were assigned to each port when they were added to the
layout. Use the Momentum Port Editor dialog box to identify which pair of ports will be treated as a differential
port.
When Momentum simulates designs containing non-consecutive port numbers, the ports are remapped to
consecutive numbers in the resulting data file. The lowest port number is remapped to 1, and remaining
numbers are remapped in consecutive order. The port numbers are not changed in the design itself. A
message in the Status window announces the change, and lists the mappings.
For example, if you are simulating a design with ports numbered 1 and 3, the following status message
informs you of the changes:
Layout has non-consecutive port numbers.

Output files will have consecutive port numbers.
layout port -> output port

1 -> 1
3 -> 2
Also, when you view results, you will see S-parameters for the differential port numbers. In the example
above, the layout would show p1, p2, p3, p4. The S-parameter results will be for combinations of the original
P1 and P3 only.

Defining a Differential Port

• Select the port that you want to assign this type
to. Note the port number.
• In the Port Editor dialog box, under Port Type,
select Differential.
• Under Polarity, make sure that Normal is
selected.
• Click Apply.
• Select the second port.
• In the Port Editor dialog box, under Port Type,
select Differential.
• Under Polarity, select Reversed.
• Under Associate with port number, enter the
number of the previously-selected port.
• Click Apply.
• Repeat these steps for other differential port pairs
in the circuit.
• Click OK to dismiss the dialog box.

Illustration of Differential Port Excitation: Direct Point Feed
direct excitationn
1.i
line feed
-1.i
line feed
ground reference
Student’s Guide A-32

Coplanar Port Properties
This type of port is used specifically for coplanar waveguide (CPW) circuits. It is similar
to a differential port, but coplanar ports are applied to objects on slot layers (that is,
where slots are used in the design). Coplanar ports should be used in situations where
an electric field is likely to build up between two ports. This can occur when:
• The two ports are close together
• Polarity between the ports develops
• The ports are connected to objects that are on slot metallization layers
• The electric field that builds up between the two ports will have an effect on the
circuit that should be taken into account during a simulation. To do this, use
coplanar ports.
Coplanar ports have the following properties:

• They can be applied to objects on slot layers only.
• They are assigned in pairs.
• Each of the two ports is excited with the same absolute potential, but with the
opposite polarity. The voltages are opposite (180 degrees out of phase). The
currents are equal but opposite in direction when the ports are on two symmetrical
lines, and the current direction is approximated for other configurations.
• The two ports must be on the same reference plane.

Coplanar Port Polarity
Be careful when assigning polarity to coplanar ports.
An incorrect choice of polarity can change the phase of
transmission type S-parameters by 180 degrees.
To verify polarity, zoom in on a coplanar port. You will

notice two sets of arrows applied to the port. One
appears when you add the port component to the
circuit. The second will appear after the mesh is
computed. It indicates the direction of the voltage over
the slot.
Note: Port numbers for CPW ports
are treated similar to the manner in
which differential ports are treated.

Defining a Coplanar Port
Note Coplanar ports can be applied to objects on slot layers only.

• Select the port that you want to assign this type to. Note the port
number.
• In the Port Editor dialog box, under Port Type, select Coplanar.
• Under Polarity, make sure that Normal is selected.
• Click Apply.
• In the Port Editor dialog box, under Port Type, select Coplanar.
• Under Polarity, select Reversed.
• Under Associate with port number, enter the number of the
previously-selected port.
• Click Apply.
• Repeat these steps for other differential port pairs in the circuit.

Coplanar Port Example: examples/Momentum/Microwave/CPW_bend_prj

Coplanar Port Example: examples/Momentum/Microwave/CPW_bend_prj

Placing and Defining Ports Note: Visualization
Coplanar Port Example displays the magnetic
currents (not the
electrical currents) for
slot metallizations.
Therefore, slots are
visualized and not metal.

Placing and Defining Ports Note: Can also be utilized for
“thick conductor” simulations
Common Mode Port Properties (more on this later)
Use common mode ports in designs where the polarity of fields is the same among two
or more ports (even modes propagate). The associated ports are excited with the same
absolute potential and are given the same port number.
Common mode ports have the following properties:

• They can be applied to objects on strip layers only
• A ground plane or other infinite metal (such as a cover) is required as part of the
design
• Two or more ports can be associated
• Associated ports are excited with the same absolute potential (and same polarity)
• The ports must be on the same reference plane
Note Port numbers for common ports are treated in the following
manner: on the layout, you will continue to see the port numbers
(instance names) that were assigned to each port when they were
added to the layout. Use the Momentum Port Editor dialog box to
identify which group of ports will be treated as a common port.
Also, when you view results, you will see S-parameters for the common
port numbers. In the example above, the layout would show p1, p2, p3.
The S-parameter results will be for combinations of P1 only.

Defining a Common Mode Port
• Select the port that you want to assign this type to. Note the port
number.
• In the Port Editor dialog box, under Port Type, select Common
Mode.
• Click Apply.
• In the Port Editor dialog box, under Port Type, select Common
Mode.
• Under Associate with port number, enter the number of the port
that you selected first. Make sure that the value in the Associate
with port number field is the same for additional ports. For
example, if you were associating three ports and the first port was
assigned as port 1, for the second and third port, the value
entered into the Associate with port number field would be 1. (For
the first port you choose, no value is entered in this field.)
• Click Apply.
• Repeat these steps for other common mode ports in the circuit.

Ground Reference Port
Ground references enable you to add explicit ground references to a circuit, which may
be necessary if implicit grounds are in your design.
Implicit ground is the potential at infinity, and it is made available to the circuit through
the closest infinite metal layer of the substrate. Implicit grounds are used with internal
ports and with single ports that are connected to objects on strip metallization layers.
There are instances where the distance between a port and its implicit ground is too
large electrically, or there are no infinite metal layers defined in the substrate. In these
cases, you need to add explicit ground references to ensure accurate simulation results.
For more information on using ground references, refer to "Simulating with Internal
Ports and Ground References" on page A-10 in the Momentum manual.
You can apply ground references to the surfaces of object. The object must be on strip
metallization layers.
Note: Multiple ground reference ports can be associated with the same port. To be
associated with a single port, the ground reference port should be a port attached to an
edge of an object in the same reference plane as the single port.

Defining a Ground Reference Port
• Select the port that you want to assign as the ground reference.
• In the Port Editor dialog box, under Port Type, select Ground Reference.
• Under Associate with port number, enter the number of the single or
internal port that you want to associate with this ground reference. Make
sure that the distance between the port and ground reference is electrically
small.
• Click Apply.

CPW with Finite Ground Planes using INTERNAL and GROUND REF ports
• Ports 1 and 2 are internal.

• Ports 3, 4, 5, and 6 are ground reference . The grounds are
associated with the internal port using the editor.

Remapping Port Numbers
Some designs contain non-consecutive port numbers. This results in simulation data files that
are difficult to use. When Momentum simulates designs containing non-consecutive port
numbers, the ports are remapped to consecutive numbers in the resulting data file. The
lowest port number is remapped to 1, and remaining numbers are remapped in consecutive
order. The port numbers are not changed in the design itself. A message in the Status
window announces the change, and lists the mappings.
For example, if you are simulating a design with ports numbered 37 and 101, the following
status message informs you of the changes:
Layout has non-consecutive port numbers.

Output files will have consecutive port numbers.
layout port -> output port
37 -> 1
101 -> 2
Port number remapping is done only for sampled and AFS CITIfiles and their corresponding
S-parameter datasets. It is not done for Visualization and far field files. The remapping is
done at the CITIfile level, and propagates to the dataset file. After remapping, all datasets
are in sync with the new port numbering.

Agilent Direct
Americas
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
United Kingdom 44 (0) 118 9276201

without notice.

5989-9597EN
Agilent EEsof EDA
Details of Momentum
• Select Mode
• Port Setup
• Mesh Generation
• Planar Solve
• Display Results

Defining Mesh Parameters
Mesh Setup Control
Global mesh is the default.
But you have choices.
In general, small
patterns are more
accurate but take
more time to solve.

Global Mesh example with Edge Mesh
1 - Port Here, the cell size is the same for
2 - Calibration Line all parts of the geometry, except for
the edges around each primitive.
3 - Mesh
4 - Edge Mesh 2 4
3
The calibration line

is automatically
drawn when the
port is defined -
more on this later.
1
NOTE: You can view the mesh, ports, and reference line
before simulating and make adjustments if desired.

Primitive Mesh example
You can combine primitive mesh,
layer mesh, and global mesh.
The center primitive of this geometry has a

different mesh (50 cells/wavelength) than the two
outside geometries (20 cells/wavelength).
1
2
Next, let’s discuss ports...

Discretion Error - Longitudinal
• Number of cells/wavelength
determines samples used for
approximation of true currents
• Typical cells/wavelength is 20
• 30 or more is fine, but will slow down
the simulation
• Minimum required to retain high-
frequency accuracy is 10
• Can retain accuracy AND speed with
10 cells/wavelength AND edge mesh
• Remember, can also have layer-specific
meshes (or even object-specific
meshes) which allows finer meshes
where needed and coarser meshes
where current density is not as high
(such as a finite ground plane)

Edge Mesh Accuracy

Mesh: Momentum versus MomentumRF
Momentum RF & Polygon Mesh
•Meshing complex geometries with POLYGONAL cells

•Eliminates “slivery” triangles
•Eliminates redundant R,L,C elements
•Uncompromised accuracy for RF frequencies 10
cells
•Strongly reduced computer memory reduction

4
cells
•Strongly reduced computation time

reduction 1
cell
mesh topology

Using Momentum Method of Moments B1(r) B2(r) B3(r)
Maxwell’s Equations I I I3
1 2
Matrix Equation ≤λ/10

• Solution process [Z].[I]=[V]
I1 L12
L1
3I2 L23 I3
• Select Mode Equivalent Circuit

L1
1C
R22 L22 L33
C22
1
[Z] = [R] + jω[L] + 1/jω [C]-1 1
• Substrate definition C12
• Port Setup
• Mesh Generation
• Planar Solve
• Display Results

The Low-frequency Breakdown Problem
• This problem is essentially one of mathematical aspect ratios. When rooftop basis
functions are used, the interaction matrix contains all of the reactances in a single
matrix. As frequency approaches zero, the inductive reactances approach zero while the
capacitive reactances approach infinity. This results in an ill-conditioned matrix.
• Any tool that uses rooftop functions as the sub-sectional basis functions will have this
problem.
• Momentum (not Momentum RF) experiences this low-frequency limitation. To help
account for this, interpolation is used for three frequencies (in addition to the selected
sweeps): DC, f0, and 2f0. The low-frequency limit (f0, typically in kHz), which is
selected in an empirical way and is a function of cell edge lengths and substrate height,
increases as cell sizes decrease (resulting in shorter edges).
• Momentum RF alleviates this problem by breaking the rooftop functions into
star and loop basis functions.

MomentumRF & Star-Loop Basis Functions
1
loop basis function star basis function
db(S11) db(S21)
Loop basis functions are solenoidal
Star basis functions are irrotational Rooftop basis functions
- give well-conditioned interaction

matrix at low frequencies
- eliminate LF breakdown of
numerical solution
db(S11) db(S21)
- give stable, accurate solutions
down to DC (both magnitude and Star-loop basis functions
phase)

Using Momentum
• Select Mode
• Port Setup
• Mesh Generation
• Planar Solve
• Display Results
More on this in the next section…

Momentum Accuracy: A couple of absolutes
• Directly simulated frequency points have –60 dB accuracy. (This noise floor was
characterized on through-lines. In other words, the observed numerical noise on
those structures is ~ -60 dB. This does not mean that valid results of < -60 dB
can not be obtained for designs with an isolation or other figure of merit that is <
-60 dB.)
• For an AFS sweep, the simulated frequency points have –60 dB accuracy while the
AFS calculated frequency points have ~ –50 to –60 dB accuracy
• The rest depends on how accurately you can define your problem.
• Here are a few benchmarks…

LTCC Filter Design
Momentum
AIR Momentum RF
Measurements
[3] 7.2 mil
[2] 3.6 mil
[1] 25.2 mil LTCC
GND
dB(S21)
7.29
mm
Momentum Momentum RF
Mesh: 20 cells/wavelength, 3 GHz Mesh: 20 cells/wavelength, 3 GHz
Frequencies: 14 Frequencies: 10
Matrix size : 218 Matrix size : 56

Process size : 14.13 MB Process size : 7.59 MB
User time : 5 m 14 s User time : 45 s phase(S21)
(*) Example from National Semiconductor

RFIC/MMIC Applications
AIR Momentum
Momentum RF
[3] 1.55 um εr=3.9
Measurements
[2] 1.7 um εr=3.9
[1] 600 um Silicon σ=12.5
GND
dB(S11)
0.30
mm 0.80
mm

Process size : 10.29 MB Process size : 3.33 MB dB(S21)
User time : 11m 09s User time : 1m 39s
PC-NT Pentium II workstation (330 MHz)

Rule of thumb: freq < 176 GHz

RFIC / MMIC Applications
Momentum
Momentum RF
Measurements
AIR
[1] 100 um GaAs
GND
0.76
mm
1.65
mm

User time : 2 m 03 s User time : 0 m 26 s
PC-NT Pentium II workstation (330 MHz) Rule of thumb: freq < 83.3 GHz

Microwave Lowpass Filter (Stripline)
Momentum
GND Momentum RF
Measurements
[2] 31 mil Duroid
[1] 31 mil Duroid

mag(S11)
GND
6.0 mm
25.4
mm
mag(S21)


Rule of thumb: freq < 5.76 GHz

RF Board Power/Ground
Momentum
Momentum RF
AIR Measurements
[1] 59 mil FR4

Momentum
GND
Process size : 20.8 MB
User time : 30 m 42 s
Momentum RF
P1 P2
50.8 Process size : 15.0 MB
mm User time : 4 m 41 s
76.2
mm
Momentum
50.8 Process size : 20.2 MB

mm
P1 P2 Momentum RF

76.2
mm PC-NT Pentium II workstation (330 MHz)

RF Board Application FYI
AIR
rectangular & triangular mesh
[1] 30 mil FR4
GND
Momentum
Mesh: 20 cells/wavelength, 1 GHz
Ports: 60
Frequencies: 6
Matrix size : 3428

35.60 mm Process size : 152.48 MB
User time : 11h 04m 51s
43.67 mm
reduced polygonal mesh
Momentum RF
Speed & Capacity
Mesh: 20 cells/wavelength, 1 GHz
memory: 3 x
Ports: 60
Frequencies: 6
Matrix size : 733

speed: 14 x
User time : 48m 24s

Packaging Application FYI
3 epoxi
4
Vchip
ref 4 ref 3
1 2
FR4
Vboard
GND
7.6 mm
port 4 ref 4
port 3 ref 3
port 2
port 1
S(1,1) S(1,2)
7.6 mm
S(1,3) S(1,4)

Process size : > 1 GB Process size : 106.57 MB
User time : > 2 days User time : 5h 17m 53s

Microwave Applications FYI
Momentum
Momentum RF
AIR
[1] 25 mil Alumina

mag(S11)
GND
mag(S21)
6.65
mm
9.90
mm
Frequencies: 18 Frequencies: 14 radiated
power
PC-NT Pentium II workstation (330 MHz) Rule of thumb: freq < 12.5 GHz

Microwave Applications FYI
Momentum
Momentum RF
[2] 185 mil AIR
[1] 25 mil Alumina

mag(S11)
GND
5.21
mm
24.82 mm
mag(S21)



Digital Application FYI
isolated trace full board
port 1
port 1
port 2
port 2
S(1,1) S(1,2) S(1,1) S(1,2)
isolated trace isolated trace full board full board
Momentum
Momentum RF

isolated trace
port 1
0.4 GHz
port 2
output
S(1,1) S(1,2)
isolated trace isolated trace

isolated trace
harmonic signal
port 1
2.33 GHz
port 2
no output
resonance
blocks the signal
S(1,1) S(1,2)
isolated trace isolated trace

harmonic signal
2.33 GHz
harmonic signal is coupled to neighboring traces

and spread around the board

Agilent Direct
Americas
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
United Kingdom 44 (0) 118 9276201

without notice.

5989-9598EN
Agilent EEsof EDA


15 minutes Break

4 hours 15 minutes
Momentum Seminar momentum_02_01 Page1

Momentum Datasets
Variables Available in the Standard Dataset
• freq Independent frequency variable
• GAMMAn Modal propagation constant of port n (calculated for single,
differential, and coplanar ports only)
• PORTZn Impedance of Port n
•S S-matrix, normalized to PORTZn
• S(i,j) S-parameters for each port pairing, normalized to PORTZn
• S_50 S-matrix, normalized to 50 ohms
• S_50(i,j) S-parameters for each port pairing, normalized to 50 ohms
• S_Z0 S-matrix, normalized to Z0
• S_Z0(i,j) S-parameters for each port pairing, normalized to Z0 of each port
• Z0n Characteristic impedance of Port n (calculated for single,
differential, and coplanar ports only, others are 50 ohms)
(Note that these are included in the datasets for Momentum simulations but not for MomentumRF)

Momentum Datasets
Variables Available in the AFS Dataset
• All standard dataset variables, plus…
• S_CONV Boolean results for AFS convergence (success=1, fail=0)

of the entire S-matrix at a given frequency
• S_CONV(i,j) Boolean results for AFS convergence (success=1, fail=0)
of S(i,j) at a given frequency
• S_ERROR Estimated error of the entire S-matrix at a given frequency

(< -60 dB for converged frequency points)
• S_ERROR(i,j) Estimated error of S(i,j) at a given frequency
(< -60 dB for converged frequency points)

Adaptive Frequency Sampling
Simple Answer to Convergence

AFS has converged unless it tells you that it hasn't converged (e.g., when the max number of points
that you specified was too low)

AFS Convergence Illustration

AFS Convergence

Momentum Datasets
Variables Available in the Far-field Dataset
• THETA Swept parameter of planar cut
• PHI Swept parameter of conical cut
• Etheta & Ephi Absolute E field strength (V) of theta and phi far-field components
• Htheta & Hphi Absolute H field strength (A) of theta and phi far-field components
• Elhp & Erhp Normalized E field strength of LHCP and RHCP far-field components
• ARcp Axial ratio, derived from LHCP and RHCP far-field components
• Eco & Ecross Normalized E field strength of co and cross polarized far-field comp
• ARlp Linear polarization axial ratio, derived from co and cross polarized
far-field components
• Gain, Directivity Gain, Directivity, Efficiency (in %), and Effective area (in m2)
Efficiency,
Effective Area
• Power Radiation intensity (in watts/steradian)

Momentum Visualization
Momentum Visualization Enables You to View and Analyze...
•Currents (surface currents)

•S-parameters (mag, re, im, phase, and dB of S(i,j))
•Transmission line data (propagation constant, characteristic impedance)
•Far-fields (radiation patterns & axial ratio in 3D and 2D)
•Antenna parameters (gain, directivity, pointing angle, etc.)

Momentum Visualization: Surface Currents
When you scroll from 0-360,
you are actually varying the
phase which illustrates the
e^jwt time dependency of
the surface currents
The lower and upper values input
into these fields represents the
lowest and highest values of the
surface current density (A/m)
which will be viewed
Note that you can also see the
effects of an edge mesh in the
You also have the current visualization (the skin
option to look at the effect is emphasized)
animated currents
when click on the
Display Properties
button
Note: when you are viewing the results for a slot metallization layer, the MAGNETIC currents are plotted instead of the
ELECTRIC currents. You will also be viewing the mesh in the slots instead of a mesh on the conductors when viewing the
mesh for a slot layer.

Momentum Visualization: Surface Currents

Momentum Visualization:
Far-field Radiation Patterns and S-parameters
Radiation Patterns are

only available with
Momentum results, not
MomentumRF

ADS Data Display: S-parameters, L, and Q of an Inductor
Powerful post processing data display allows you to take
advantage of countless built-in functions and provides the
flexibility to wrote your own (through both measurement
equations in a schematic or equations in a data display page).

Agilent Direct
Americas
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
United Kingdom 44 (0) 118 9276201

without notice.

5989-9599EN
Agilent EEsof EDA
Detailed Presentation on Momentum Advanced

Topics - (Part 1 of 5)


15 minutes Break

4 hours 15 minutes

Momentum Component (EM/circuit co-simulation)
• EM/Circuit co-simulation from the

schematic environment
• Transparent integration of electromagnetic
simulators at the schematic design level
Layout setup
• Include physical layout parasitics in Momentum Component Generation
schematic
• Momentum simulation options accessible
from schematic
• Compiled Layout Components listed in
project’s hierarchy ADS circuit simulation
• Model database for reuse option
• ADS 2002C: EM/Circuit co-
optimization

Example included in ADS 2002 & higher
C:\ADS2002\Examples\Momentum\emcktcosim\LTCC_prj

C:\ADS2002\Examples\Momentum\emcktcosim\LNAEmCktCosim_prj

• Example begins with schematic of LNA
and uses Layout>Generate/Update
Layout to create artwork
2. Note that vendor component libraries

were utilized for lumped element and
active device artwork. Also note that
a ground plane has been added with
uniform clearance around
traces/components.

3. A symbol is defined for the schematic

subcircuit and then it is placed in a top
level design for simulation using the
component library browser. (The
results of this simulation will be
compared to the results of the next
simulation, which will include the
physical effects.)

Example included in ADS 2002 & higher 4. The artwork is now modified to
include the vias (to connect the
“poured” ground plane to the explicit
ground plane) and ports are placed at
all connection points (input, output,
bias, and on each node of the lumped
elements)

Momentum Component (EM/circuit co-simulation) Layout/Momentum
Example included in ADS 2002 & higher (slightly modified) Component parameters
will be discussed more
later (co-optimization)
5. The standard example is then slightly

modified to include layout component
parameters (new in ADS 2002C).
These parameters will be used to see
the effects of a via location on gain.
Modified version provided:
(LNAEmCktCosim_prj.zap)
Example included in ADS 2002 & higher (slightly modified)
6. The Layout/Momentum component is

then created

7. Next, the Layout/Momentum component is placed in a schematic using the component library browser (just
like any other subcircuit/component). All of the lumped elements and the active device are then connected to
the pins (ports in layout are replaced with pins in the Momentum Component symbol).

8. Once the model details are selected (Mode – MomMW, MomRF, or data file; Frequency range; Mesh
properties), the parameters of the Layout/Momentum component are then defined to be variables, which will
be passed down from the top design. This is made possible by the next step, which uses the File>Design
Parameters submenu.
9. The variables are now defined for this subcircuit. Note that we could have just placed the Layout/Momentum
component directly into the top level schematic, but this illustrates two methods of parameterization in a
schematic.

10. The subcircuit that includes the

Layout/Momentum component is then
placed in a top level design for
simulation using the component
library browser. (The results of this
simulation will be compared to the
11. Note the use of variables which are
results of the original simulation,
swept using the Parameter Sweep
which did not include the physical
component.
effects.)
12. When each iteration of the S-parameter

simulation encounters the
Layout/Momentum component with
new values for parameters, a
Momentum simulation is also
Modified version provided: automatically invoked in the
(LNAEmCktCosim_prj.zap) background (for each case that was
not previously solved).

13. Finally, the results from each approach

are compared. Surface currents for
the Momentum component (layout) can
also be studied using Visualization.
Initial schematic simulation
Momentum component in schematic
Momentum component in schematic
Modified version provided: with swept variable for via location
(LNAEmCktCosim_prj.zap)

Momentum Co-Optimization
EM/circuit co-optimization (layout/Momentum components with parameters)
Adding Layout Parameters NEW in ADS 2002C

Enables to sweep, tune or optimize geometrical variations in the
layout
- typical dimensions (length, width, gaps, spacing,…)
- interdependent layout modifications (e.g. length and width varying simultaneously)
- port locations
Two ways to create a parameterized layout component

1. Using nominal/perturbed layout artwork (Momentum Optimization)
“Nominal/Perturbed” layout parameter

2. Using existing (built-in or GCC defined) layout artwork macro’s
“Subnetwork” layout parameter

Adding Layout Parameters
Momentum > Component > Parameters
• Opens the “Layout Component Parameters” dialog

Defining a Nominal/Perturbed Layout Parameter
Define the name of the parameter

Define the type of parameter
Enter the nominal value
Enter the perturbed value
Edit the perturbation
if no AEL artwork macro available

Steps
1. Select points in layout

2. Select perturbation type
3. Insert perturbation values
4. Apply the perturbation
Repeat these steps

Click OK to terminate
AEL artwork macro is created

primitive layout component

Adding Layout Parameters to a Subnetwork
Defining a Subnetwork Layout Parameter
Define the name of the parameter

Define the type of parameter
Enter the default value
Associate the parameter with a
subnetwork parameter
If artwork macro IS available

hierarchical layout component
Adding Layout Parameters to a Subnetwork
Use the subnetwork layout parameter to set the parameter values of one or
more subnetwork parameters in the design

Creating a Component
Momentum > Component > Create/Update
• Opens the “Create Momentum Component” dialog
• Dialog Entries: Symbol, Model Parameters and Model Database
These parameters are a

subset of the
Momentum simulation
control options in the
Layout Environment and
can be set from within
the Schematic
Environment

EM Model Database
The simulated S-parameter models are

stored in an EM Model Database for later
reuse
During the Component Create/Update, the

user has the option to:
- delete all previous entries in the model

database
- add the last simulation results obtained

from Momentum simulation in Layout to
the model database

Instance Parameter Dialog
Double clicking on the Layout Component Instances in the Schematic Design
Environment opens the Instance Parameter Dialog
In the Model Page, the user can specify the

- Model Type selection
- Model Parameter values
- Model Database Reuse option
In the Parameters Page, the user can

specify the layout parameter values and
the optimization setup (optional)
In the Display Page, the user can specify

which model parameters are visible in the
schematic design

Model Interpolation
Instance Dialog box

The EM model database can use
interpolation to significantly enhance
the efficiency of the co-simulation
The Layout Parameters are treated as

continuous parameters
Pushing the ‘Options..’ button brings up

the ‘Set Interpolation Options’ dialog
Allows to specify the interpolation delta

values for each layout parameter
Default values for the interpolation deltas are

provided (derived from the model parameters)

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and
higher) examples/Momentum/emcktcosim/LTCC_prj

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and

Electronic
notebook for
LTCC
EM/circuit co-
optimization
example
(shipped with
ADS 2002C and
examples/Momentum/emcktcosim/LTCC_prj
higher)

Agilent Direct
Americas
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
United Kingdom 44 (0) 118 9276201

without notice.

5989-9600EN
Agilent EEsof EDA



15 minutes Break

4 hours 15 minutes

An Aside: Vias in Momentum
1. Polygon
Vias must be mapped
2. Rectangle or square through each individual
substrate layer they pass
3. Polyline through
• For vias, only vertical currents and surface impedances are taken into account (for now). Keep in
mind that the horizontal and rotational currents are not included. One possible “trick” that might
be used to obtain more complete current calculations is to break up the via structure (could be any
shape, even that of a transmission line or spiral inductor) into a few thinner layers and include the
geometry on horizontal metallization layers as well. Make sure that provide the conductivity and
thickness parameters for only one of the horizontal metallization layers; otherwise, excess loss
will be calculated.

An Aside: Vias in Momentum
• Vias are treated as one cell (in the vertical axis);

therefore, the thickness of substrates which
contain vias is limited to ~ 1/10th to 1/20th of
a wavelength (at the highest simulation
frequency). Vias passing through thicker
substrates can be accurately represented by
splitting these thick substrates into multiple
layers.
• Please note that you can view both the mesh

and the surface current density (A/m) of via
structures using Momentum Visualization

Thick Conductor Simulations
Momentum model Finite Thickness Conductors
• Zero thickness approach

• loss formulation
• Finite thickness approach
• loss formulation
• current modeling

Conductor Loss in Momentum: Zero Thickness Approach
• The Surface Impedance Concept is used to model

conductor losses in metallizations
σ Zs(t,σ,ω)
t
3D conductor Sheet conductor
• The Surface Impedance Model Zs(t,σ,ω) takes the finite

thickness and frequency dependency (skin effect) of the
conductor loss into account

Conductor Loss in Momentum: Zero Thickness Approach
• The Surface Impedance formula:

jωµ (1 + j ) æ (1 + j ) t ö
Zs = coth (γ c t ) σ>>ωε Zs = coth çç
σ + jωε σ δs è δs
2
γc = jωµ (σ + jωε ) δs =
ωµσ
1
LF : Zs = σ t LF currents run in entire cross
σt section of the metalization
(1 + j ) HF currents run in SINGLE skin

HF : Zs = σ δs
depth surface layer
σ δs
Skin depth 1 MHz 75 µm
σ=4.5e7 S/m 10MHz 23.7 µm
100MHz 7.5 µm
1 GHz 2.37 µm

Conductor Loss in Momentum: Finite Thickness Approach
3D conductors:
drawn on 2 layers (σ, t/2) connected with via’s

LF : σ t/2
LF currents run in entire cross

section of the metalization Zs(t/2,σ,ω) for top and
bottom metalization layer
HF : σ δs
HF currents run in DOUBLE

skin depth surface layer
Better loss modeling

Better current modeling (inductive)
Current Modeling in Momentum for 3D Conductors
z
y
x
x,y surface currents on top and bottom of finite thickness conductor
z-surface currents on vias (side walls of finite thickness conductor)

Conductor Loss in Momentum - what to use?
w
Rule of thumb:
t
h
ground plane
w/t > 5
h/w > 10 Use 1 zero thickness conductor with correct loss specification (thickness, conductivity)
σ, t
σ, t/2 strip
other cases via
strip
2 metallization layers + vias σ, t/2

Thick-conductor Approach: Two Approaches
“layer3”: perfect conductor “layer2”: σ, t/2
thick conductor 2 metallization layers + vias
σ, t/2 strip
via Air (E0) t
strip
σ, t/2 Substrate (Er)
h
“layer1”: σ, t/2
Port 1 Port 1 Port 3 …

Port 2 Port 2 Port 4
For Single Trace Stimulus: For Two or More Coupled Traces:

• Common-mode ports • Single ports in layout, recombine as common-mode in schematic

Thick-conductor Approach: For Single Trace Stimulus
Port 4
Port 2
Example:
• Ports 1&3 are associated as Port 3
common-mode ports, as are
ports 2&4 Port 1
For Single Trace Stimulus:

• Common-mode ports in Layout/Momentum are fine



15 minutes Break

4 hours 15 minutes

Microstrip Coupled-line Filter Example: initial layout for filter generated
from schematic (substrate definition also updated from schematic)
filter_thick_metal_prj/DA_CLFilter1_untitled1.dsn

Microstrip Coupled-line Filter Example: verify substrate and
metallization definitions

Microstrip Coupled-line Filter Example: initial MomentumRF simulation

Microstrip Coupled-line Filter Example: save as new design and flatten
components to enable copy-to-layer for the thick-conductor method
filter_thick_metal_prj/DA_CLFilter1_untitled1_thick.dsn

Microstrip Coupled-line Filter Example: now begin copying the artwork
to the first two additional layers

Microstrip Coupled-line Filter Example: now copy the artwork to the
second of two additional layers

Microstrip Coupled-line Filter Example: modify substrate definitions to
include another “substrate” layer that is the thickness of the metal

Microstrip Coupled-line Filter Example: map one of the copied artwork
layers as a strip conductor above the new “conductor substrate” and
the other as a via in the new “conductor substrate”
Define the top and bottom conductors with

the proper conductivity and ONE-HALF
the total conductor thickness.

Microstrip Coupled-line Filter Example: place additional port on the
newly mapped strip conductor layer at each original port location

Microstrip Coupled-line Filter Example: assign and associate the
overlapping ports as common-mode ports

Microstrip Coupled-line Filter Example: define frequency sweep and
simulate

Microstrip Coupled-line Filter Example: …something to REMEMBER!
REMEMBER:
Vias may not coincide with a port or touch a port.
So how is this simulation accomplished?

Microstrip Coupled-line Filter Example
By slightly moving the coincident edge of the via layer
so that it no longer touches the port!
Step 1. Begin by making the via layer the only selectable layer

Step 2. Drag a selection box around the edge to be moved

Step 3. Select the “Move Relative” tool

Step 4. Enter a sufficiently small (but non-zero) distance to move the

edge so that it does not exactly overlap the port.

Step 5. Zoom in to verify that the edge was indeed moved. Then repeat this for any other via edges that
coincided with a port. Now it is safe to simulate without errors.

Microstrip Coupled-line Filter Example: compare results for different approaches
filter_thick_metal_prj/compare_approaches.dds

filter_thick_metal_prj/compare_approaches.dds Note that for cases that do not include coupled port
stimuli, Momentum common-mode ports yield
correct results. This can be verified by re-defining
these ports as Single ports and then recombining
the ports as common-mode in a schematic. The
results are compared below (and are identical). We
will see that only the latter approach is valid for
cases that include coupled ports.

Thick-conductor Approach: For Coupled Traces
Layout Port 6 Port 5 Port 4
Schematic
Term 6 Term 5 Term 4
Term Term Term

Term1 Term1 Term1
Num=1 Num=1 Num=1
Z=50 Ohm Z=50 Ohm Z=50 Ohm
Example:
• Ports 1-12 are Single ports
Term Term Term

Term1 Term1 Term1
Num=1 Num=1 Num=1
Term 1 Z=50 Ohm

Term 2 Z=50 Ohm
Term 3 Z=50 Ohm
For Two or More Coupled Traces:

• Use Single ports in the Layout Momentum Simulation
• Recombine these results as common-mode ports in a schematic (this
will ensure port calibration for the coupled ports)

(When the Momentum Dataset is Available – Method 1)
p_6 p_5 p_4
S-PARAMETERS
p_1
p_2
1
12 11 10
9
p_3
p_2
S_Param Using an SnP data item
SP1
(S12P in this example)
2 8
p_3 3 Ref 7
p_1 Start=
4 5 6
Stop=
Step=
S12P p_4 p_5 p_6
SNP1
File="coupled_lines_thick_momRF_50_single_ports.ds"
p_1 p_2 p_3 p_4 p_5 p_6

T erm Term Term Term T erm T erm
T erm1 Term2 Term3 Term4 T erm5 T erm6
Num=1 Num=2 Num=3 Num=4 Num=5 Num=6
Z=50 Ohm Z=50 Ohm Z=50 Ohm Z=50 Ohm Z=50 Ohm Z=50 Ohm
OR
Tips/Tricks
• To reduce wiring confusion,

use wire/pin label tool to
• Place an S2P • Use on- make connections
Data Item screen • Use S12P data
editing to item to access
change S2P dataset from
to S12P Momentum
(When the Momentum Dataset is Available – Method 2)
Lead_frame_package_thick_momRF_50_single_ports_component
Lead_frame_package_thick_momRF_50_single_ports_component_1
ModelType=Dataset
ModelFile=".\data\Lead_frame_package_thick_momRF_50_single_ports.ds"
Using a Layout/Momentum
p_1 p_6
Component as a data item
p_2 p_5
Be sure to select “Dataset” as the ModelType
(not “Momentum MW” or “Momentum RF”) and then
browse to the Momentum dataset that uses all
p_3 p_4
Single ports
Ref
p_1 p_2 p_3 p_4 p_5 p_6

Term Term Term Term Term Term S-PARAMETERS
Term1 Term2 Term3 Term4 Term5 Term6 S_Param
Num=1 Num=2 Num=3 Num=4 Num=5 Num=6 SP1
Z=50 Ohm Z=50 Ohm Z=50 Ohm Z=50 Ohm Z=50 Ohm Z=50 Ohm Start=
Stop=
Step=
Tips/Tricks Note that since the 12 layout ports in this

example need to be reconnected as 6
common-mode ports, it is acceptable (even
desirable) for the pin connections to short
together each pair of pins in this component
• Use the component library • To reduce wiring confusion,
symbol. (It is impossible not to since the pins
browser to place the use wire/pin label tool to
co-exist from this top view of the layout.)
Momentum Component make connections
Comments on Overlapping (Bird’s Eye View)
Ports in Momentum Layout Components
When would the pins in the Layout Component symbol overlap/short?

• When common-mode ports overlap on different layout layers (same xy-coordinates,
different z-coordinates). You usually want these ports to be connected/shorted in the
schematic, as shown in the previous slide.
• When internal ports and ground reference ports overlap on different layers (same xy-
coordinates, different z-coordinates). You usually DO NOT want the pins for internal
ports to overlap/short pins that represent the ground reference ports in the schematic.

Comments on Overlapping (Bird’s Eye View)
Ports in Momentum Layout Components
(Workarounds for cases when you DO NOT want the pins for internal
ports to overlap/short pins for ground references in the schematic)
How do I avoid shorting these pins together?
• When the ports are attached to an edge, it doesn't matter WHERE on that edge the port is attached, because
the current is injected in a distributed manner along the edge. So, shifting the ground reference port(s)
slightly won't make any difference in the results. The Layout Component symbol will now reflect non-
overlapping pins for the internal and ground reference ports.
• When the ports are point injection ports (in the middle of the metal) then shifting them WILL make a
difference in results, so be careful. So what are your options when this is the case?
• Edit the symbol page of a layout component symbol! When the Layout component is created, edit its
symbol page and look for the port pins. You can move them a little in order to correctly use the ones
that are on top of each other. This will not change the results. I have tried this and it works! You just
need to be sure to remember which pin is which, or label them with text in the symbol.
• Use the "black box" symbol, then all is fine. This would be less work than the prior suggestion, but it
would not be as "pretty“.
The next example will also demonstrate this…
Agilent Direct
Americas
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
United Kingdom 44 (0) 118 9276201

without notice.

5989-9601EN
Agilent EEsof EDA



15 minutes Break

4 hours 15 minutes

Thick Conductor Simulations coupled_lines_prj/mlayer_lib.dsn
Coupled Lines Example: starting in a schematic

Coupled Lines Example: Layout>Generate/Update Layout

Coupled Lines Example: Auto generated layout

Coupled Lines Example: Update substrate definition from schematic

Coupled Lines Example: verify substrate and map metal layers

Coupled Lines Example: edit/assign port types

Coupled Lines Example: flatten component

Thick Conductor Simulations coupled_lines_prj/coupled_lines_momRF_50.dsn
Coupled Lines Example: save as new design

Coupled Lines Example: setup mesh

Coupled Lines Example: setup simulation (zero-

thickness conductor standard approach)

Thick Conductor Simulations coupled_lines_prj/coupled_lines_thick_momRF_50.dsn
Coupled Lines Example: save as new name to be used for

thick conductor approach

Coupled Lines Example: make an empty layer insertable

Coupled Lines Example: select original geometries

Coupled Lines Example: copy original geometries to the empty layer

Coupled Lines Example: make another empty layer insertable

Coupled Lines Example: copy geometries to this third layout layer

Coupled Lines Example: modify substrate to include a “dummy” layer for

the metal thickness and map the new (copied) via and strip layers
Note: the conductor

thickness has been split
between the top and
bottom strip layers that
form the new thick
conductor shell.

Coupled Lines Example: insert ports on the newly mapped top

metallization layer

Coupled Lines Example: assign/associate ports as common mode

Coupled Lines Example: setup the new simulation

Thick Conductor Simulations coupled_lines_prj/coupled_lines_thick_momRF_50_single_ports.dsn
Coupled Lines Example: save as new name to be used for

New thick conductor approach

Coupled Lines Example: re-define ports as Single ports

Coupled Lines Example: setup the new simulation

Thick Conductor Simulations coupled_lines_prj/coupled_lines_thick_momRF_50_single_ports_schematic.dsn
Coupled Lines Example: now, recombine single port data as common-mode
Accomplished by either using an SnP data item…

Thick Conductor Simulations coupled_lines_prj/coupled_lines_thick_momRF_50_single_ports_component_schematic.dsn
Coupled Lines Example: now, recombine single port data as common-mode
NOTE: IF the version of ADS that

you are using shows unconnected
pins (red diamond around pin)
when using this Layout
Component method, please edit
the Layout Component symbol so
that you can see both pins for
those that overlap. If the
connection is good (blue), then no
need to worry about this.
…or by using a Layout/Momentum Component as a data item

Thick Conductor Simulations coupled_lines_prj/coupled_lines_thick_momRF_50_single_ports_component_schematic.dsn
Coupled Lines Example: re-defining Layout Component symbol
Step 1. Push into symbol in

schematic

Thick Conductor Simulations coupled_lines_prj/coupled_lines_thick_momRF_50_single_ports_component.dsn
Step 2. Enable symbol view

Thick Conductor Simulations coupled_lines_prj/coupled_lines_thick_momRF_50_single_ports_component.dsn
Step 3. Separate
overlapping pins by moving
one of each pair.

Thick Conductor Simulations coupled_lines_prj/coupled_lines_thick_momRF_50_single_ports_component_test.dsn
Step 4. After separating all

overlapping pins, save the
design. Note that the
specific location of each pin
is inconsequential since this
is only the symbol, not the
actual artwork!

Thick Conductor Simulations coupled_lines_prj/coupled_lines_thick_momRF_50_single_ports_component_schematic_test.dsn
Step 5. Now you can

successfully wire all pins.
(No more unconnected pins!)

Thick Conductor Simulations coupled_lines_prj/compare_approaches.dds
Coupled Lines Example: open a data display to compare results
Please note the following details regarding this data display:
• This data display contains 3 pages (return loss, insertion loss, and coupling
data)
• There are actually 5 approaches being compared (which are all included in this
coupled_lines_prj project). The 5 designs are:
• coupled_lines_momRF_50.dsn standard MomRF simulation using Single ports
• coupled_lines_thick_momRF_50.dsn thick-metal simulation using common-mode ports
• coupled_lines_thick_momRF_50_single.dsn thick-metal simulation using our legacy
approach with only one Single port per node
• coupled_lines_thick_momRF_50_single_ports_schematic.dsn schematic re-combination
of MomRF simulation using TWO Single ports per node (using SnP data item) Approach #4 and approach #5
• coupled_lines_thick_momRF_50_single_ports_component_schematic_test.dsn schematic are equivalent AND are the best
re-combination of MomRF simulation using TWO Single ports per node (using Layout method for this type of
Component as a data item) simulation
Coupled Lines Example: comparing return loss (log magnitude & phase)

Coupled Lines Example: comparing insertion loss (log magnitude & phase)

Coupled Lines Example: comparing coupling (log magnitude)

Coupled Lines Example: visualizing various approaches



15 minutes Break

4 hours 15 minutes

Spiral Inductor Simulations
Spiral Inductor simulation will be shown with the Advanced Model
Composer Lab
Ref
The project LTCC_example_AMC_prj illustrates co-simulation, swept parameters of a Layout component, co-
optimization, and Advanced Model Composer capabilities for a multilayer LTCC spiral inductor. The
inductance (L) and quality factor (Q) are spiral_param
calculated for several parameter values (length of sides, length of
input line, length of output line). spiral_param_1
This highlights the usage of nominal/perturbed parameters.
An additional example is included for reference. The archived ADS

project is entitled AMC_lab1_prj.zap and is discussed in detail (with
specific steps to perform) in the presentation 2003A_physical_design.ppt.
This highlights the usage of subnetwork parameters.

Agilent Direct
Americas
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
United Kingdom 44 (0) 118 9276201

without notice.

5989-9602EN
Agilent EEsof EDA



15 minutes Break

4 hours 15 minutes

Advanced Model Composer (new in ADS 2003A)
• Previous ADS Capabilities • New in ADS 2003A

• Model Composer – ADS 2001 • Advanced Model Composer
• Fast & accurate circuit-level • Model Composer capabilities
passive parameterized models plus:
created w/ Momentum EM
• Arbitrary user-defined
technology
parameterized shapes
• Automated model generation
• Passive components including
• Patented MAPS technology spiral inductors, TFR, etc.
• Custom standard-shape • No end-user license required for
microstrip & stripline libraries use of model in schematic
and substrates
• Model generation uses AMC and
• Generates model, symbol & Momentum license
layout
• ADS Design Kit format
• Provided with Linear license


• Integrated Momentum EM Simulator • Advanced Model Composer
• Momentum RF • Arbitrary shapes
• Model Composer
• Parameterized
• Co-Simulation w/Layout Components
• Creates Design Kit
• Co-Optimization w/Layout Components
• Automated
30 GHz transition • Fast Simulation
• No end-user license required
• Momentum
• Queuing/Batch processing

• Generate libraries for wide array of custom shaped parameterized passive
models, including:
• Spiral inductors
• Thin Film Resistors
• MIM capacitors
• Arbitrary shaped
matching networks
MomentumEM
Momentum EMengine
engine
modelingsource
modeling source
MAPSpatented
MAPS patentedtechnique
technique
forcompact
for compactmodel
modelfitting
fitting
GeneratedLibraries
Generated Librariesdon’t
don’t
requireusage
require usagelicense
license


15 minutes Break

4 hours 15 minutes

Momentum Co-Simulation, Co-Optimization, and AMC
An LTCC Spiral example that utilizes all of these capabilities!
Note that the starting point for this

design was the spiral.dsn included in
LTCC_example_AMC_prj/spiral_param.dsn
the LTCC emcktcosim project.
An LTCC Spiral example: baseline design to which we will add parameters

An LTCC Spiral example: add first parameter for the input line length
Note that the perturbation will need to have a

magnitude of 50 um

An LTCC Spiral example: select the objects to be moved and enter the
perturbation type and value
LTCC_example_AMC_prj/spiral_param_length_in.dsn

An LTCC Spiral example: click OK to return to main design
LTCC_example_AMC_prj/spiral_param_length_in.dsn

An LTCC Spiral example: now add parameter for output line length
Note that the perturbation will need to have a

magnitude of 50 um

An LTCC Spiral example : select the objects to be moved and enter the
LTCC_example_AMC_prj/spiral_param_length_out.dsn

LTCC_example_AMC_prj/spiral_param_length_out.dsn

An LTCC Spiral example: now add parameter for length of spiral sides
Note that each perturbation will need to have a

total magnitude of 100 um

“Stretch” (shrink) objects on left side of spiral
LTCC_example_AMC_prj/spiral_param_length_sides.dsn

“Stretch” (shrink) objects on top side of spiral

“Stretch” (shrink) objects on right side of spiral

“Stretch” (shrink) objects on bottom side of spiral

All sides are now 100 um shorter

An LTCC Spiral example: click OK to accept parameters

An LTCC Spiral example: create a layout component symbol for schematic use

An LTCC Spiral example: click OK to create Layout Component symbol

An LTCC Spiral example: use component library browser to locate a place an
instance of the Layout Component in a new schematic
LTCC_example_AMC_prj/spiral_param_schematic.dsn

An LTCC Spiral example: Layout Component placed, now setup simulation

An LTCC Spiral example: double-click Layout Component to verify model details

An LTCC Spiral example: now verify values of parameters (0 nominal value)

An LTCC Spiral example: baseline results for the Inductor
LTCC_example_AMC_prj/spiral_param_schematic.dds
(page “L and Q”)
An LTCC Spiral example: baseline results for the Inductor (L & Q only)
(page “L and Q only”)
An LTCC Spiral example: setup a swept parameter simulation
LTCC_example_AMC_prj/spiral_param_schematic_sweep.dsn

An LTCC Spiral example: view resulting L & Q for each parameter sweep
(page “L and Q only (swept params)”)
An LTCC Spiral example: re-use swept cases to optimize for desired L & Q
LTCC_example_AMC_prj/spiral_param_schematic_optimize.dsn

An LTCC Spiral example: optimized L & Q (and resulting parameter values)
(page “L and Q only (optimize)”)
An LTCC Spiral example: create Advanced Model Composer model

An LTCC Spiral example: define specific sweep types for model parameters

An LTCC Spiral example: define specific sweep types for model parameters

An LTCC Spiral example: during model generation, can view status at any time

An LTCC Spiral example: can see status of current simulation, too

An LTCC Spiral example: final status – model generation complete
An LTCC Spiral example: create design kit
An LTCC Spiral example: install design kit to access the AMC model

An LTCC Spiral example: selecting the design kit files and installation level

An LTCC Spiral example: successful installation of kit

An LTCC Spiral example: use component library browser again to find AMC
model from installed design kit
LTCC_example_AMC_prj/spiral_param_schematic_sweep_AMC.dsn

An LTCC Spiral example: select DK AMC model from Layout Components library

An LTCC Spiral example: note that parameters are just as any standard
component, now (not a Momentum Layout Component) – define values

An LTCC Spiral example: compare swept parameter results to those from
Momentum Layout Component results
LTCC_example_AMC_prj/spiral_param_schematic.dds (page “L and Q only (swept params, compare co-sim vs AMC)”)

An LTCC Spiral example: optimize for L & Q using AMC model
LTCC_example_AMC_prj/spiral_param_schematic_optimize_AMC.dsn

An LTCC Spiral example: compare results (desired L & Q values obtained with both AMC and
Momentum Layout Component – if want same optimum parameter values, would need to
lock values of 2 params and optimize 1)
LTCC_example_AMC_prj/spiral_param_schematic.dds (page “L and Q only (optimize, compare co-sim vs AMC)”)

Momentum Co-Optimization/Tuning
EM/circuit TUNING (layout/Momentum components with parameters)
Example: Tuning of layout parameters
microstrip coupled stubs
L1 2 layout parameters: L1 and L2
L2
Step 1. Parameter Sweep variable L1 = L2 = L mil
to fill the EM model database sweep L = 100 … 130 mil step 3 mil
L2 [mil] m000
130
Coupled_stubs 100
example
(shipped with 100 130 L1 [mil]
ADS 2002C and
higher) examples/Momentum/emcktcosim/Coupled_Stubs_tune_prj

Example: Tuning of layout parameter
Step 2. Set the Interpolation Delta equal to or greater than the distance
between the samples
L1-step = 3 mil
L2-step = 3 mil
Delta = 6 mil
Coupled_stubs
example
(shipped with
ADS 2002C and
higher) examples/Momentum/emcktcosim/Coupled_Stubs_tune_prj

Example: Tuning of layout parameter
Step 3. Start the tuning session
Coupled_stubs
example
(shipped with
ADS 2002C and
higher)

Note: Momentum Batch Mode Queue Manager
New in
ADS 2003A
Batch Mode Simulation

Batch mode simulations are setup and controlled using the Queue Manager dialog box. The Queue Manager is automatically started when
submitting a queued job, or it can be invoked by selecting Tools > Queue manager in the ADS Main window. Exiting ADS will stop the
Momentum queue and all running jobs.
When the queue is running in normal operation (queue connected), the first waiting job will automatically start after the current job
finishes. If the queue is in Queue Disconnected mode, then the first waiting job will not start automatically when the current job finishes.
When connected, the Queue Manager can be used to view jobs in the queue. When the queue is disconnected, or hasn't been started, the
Queue Manager not only can be used to view jobs in the queue but can also be used to add jobs, modify or delete scheduled jobs, or
change the order of jobs in the queue.
Note Jobs cannot be modified using the Queue Manager when they are currently active (simulating). To modify these jobs, disconnect
the queue, stop the simulation, (see Stopping a Simulation) make your modifications, and resume.
To run a batch mode simulation:

•Define the design you want to simulate.
•In the Momentum > Simulation > S- parameters dialog box, select Queued in the Process Mode and then select simulate.
•Repeat the previous two steps to add additional simulations to the queue.
•Select Start Simulations in the Queue Manager to begin simulating after all simulations are queued.
Note If the Queue Manager indicates that simulations are running when they are not, type queue_reset() in the command line
dialog box to fix the problem.
Caution Running Momentum simulations in parallel (e.g., a direct Momentum simulation and a queued Momentum simulation)
from the same ADS session may cause the loss of simulation data if both simulations end simultaneously
Note When Momentum batch simulations are run from different projects, the active project directory changes when the
simulation finishes if a new one needs to be loaded from a different project.

Note: Momentum Batch Mode (Legacy Method)
This is discussed in great detail (with an included example project,
L_optimize_using_artwork_macro_prj.zap) in the appendix
provided in the seminar CD. This method also includes a legacy
parametric design capability.

Agilent Direct
Americas
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
United Kingdom 44 (0) 118 9276201

without notice.

5989-9603EN
Agilent EEsof EDA



15 minutes Break

4 hours 15 minutes

Extras – What’s in the Folder?
Momentum Box/Waveguide Information
• This is a topic for which some advanced users might want to
know the theory behind our implementation.
• For a thorough discussion on how the standard waveguide and
box models are implemented, see the document:
“Box_waveguide.pdf”

Momentum Configuration Information
Momentum config file.txt

RLGC Netlist from MomentumRF Information
MomentumRF_RLGC_equivalent_circuit_netlist.ppt

Material Properties “DesignGuide” Materials.deb



Remember…
Momentum Examples
Included with ADS

Remember… Links to Additional Momentum Resources
Momentum Product Webpage: http://eesof.tm.agilent.com/products/e8921a-a.html
Momentum Technical Articles Webpage: http://eesof.tm.agilent.com/products/e8921a-b.html#Technical Articles
Co-simulation with Momentum Layout Components

(online demo): http://eesof.tm.agilent.com/demos/#usability
Free MomentumRF NetSeminar (from 2001): http://netseminar.com/nss/showSeminar?sem_num=413
Advanced Model Composer Product Webpage: http://eesof.tm.agilent.com/products/e8926a-a.html
AMC Technical Articles Webpage: http://eesof.tm.agilent.com/products/e8926a-b.html#Technical Articles
Momentum Training Class Webpage: http://eesof.tm.agilent.com/products/e8921a-d.html#Training Classes

(also available as eTraining class): http://eesof.tm.agilent.com/education/etraining.html
Momentum Support Docs on Knowledge Center:

http://edasupportweb.soco.agilent.com:7778/portal/page?_pageid=36,34228&_dad=portal&_schema=PORTAL&cid=53737&a=1&lang=1
Momentum Support Examples on Knowledge Center:

http://edasupportweb.soco.agilent.com:7778/portal/page?_pageid=36,34228&_dad=portal&_schema=PORTAL&cid=8895
Momentum Discussion Forum on Knowledge Center:

http://edasupportweb.soco.agilent.com:7778/cgi-bin/yabb/YaBB.pl?board=ads_momentum
Momentum Manual (online version): http://eesof.tm.agilent.com/docs/adsdoc2002C/mom/index.html
Knowledge Center: http://www.agilent.com/find/eesof-knowledgecenter
Tech Support: http://www.agilent.com/find/eesof-support

Wrap-up: Q&A
Questions?

Agilent Direct
Americas
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
United Kingdom 44 (0) 118 9276201

without notice.

5989-9604EN
Agilent EEsof EDA
Momentum Appendix
APPENDIX
• Graphical Cell Compiler (GCC)
• Momentum Batch Simulations
• A Method for Utilizing Momentum Batch

Simulation Results in EM/circuit Co-
optimizations (useful before true co-optimization
was introduced)
• Momentum Optimization (a review of the

original capability)
Momentum Seminar Page 1

Note: Momentum CAN be run in a batch mode
• Obtain “usermenu.ael” and instructions from your local ADS Applications Engineer (AE).
• Once you load “usermenu.ael” in your $HOME/hpeesof/de/ael directory (this is usually the
users/default/hpeesof/de/ael directory), restart ADS and you will see a new user menu in the layout window
that enables you to run Momentum in a batch mode.
• You can even sweep the parameters of a parameterized design.

Momentum Optimization using Batch Simulation
First Step : Initial Layout entry (example using Graphical Cell Compiler)
This GCC macro was taken from

\ADS2003A\Examples\RF_Board
\GCC_examples_prj

Custom Layout Macros: Graphical Cell Compiler FYI
ADS Components can call custom layout macros
Allows for Parameter-driven geometry creation
Layout Macros are AEL functions
Graphical Cell Compiler can be used to generate layout macros
Writing your own macros allows for extra flexibility:

-ports placed on specific layers
-rounding functions to keep ports on-grid
-custom warning or error messages
-physical dimensions calculated from electrical data
-Boolean flags to control layout - clockwise, counter-clockwise, etc.

Graphical Cell Compiler (Macro): Introduction FYI
• Purpose:
• adding Parameterized Artwork Macros (PAM)
• Advantages:
• simpler than coding in AEL
• getting started very quickly
• create a special model quickly, without the need to know AEL

Graphical Cell Compiler FYI
Macro control operations:

●Stretch / Move
●Flip / Rotate
●Repeat
●Polar
Viewer
●Modify
●Add
●Delete
●Detail
Compile

Graphical Cell Compiler FYI
Stretch Flip / Rotate

stretch = - 25 mil
rotation = 33
rotation = 90
original original
stretch = 150 mil rotation = 0
Repeat Polar
original
original polygon
rectangle path
repeat x and y = 5 x 4 circle polar sweep PI radians
polyline

GCC Example FYI
C:\ADS2003A\Examples\RF_Board\GCC_examples_prj
start with Parameterized Inductor that is a

function of :
turns
space
width
sides
becomes

GCC: How to build a PAM FYI
1) Define the artwork graphically in a Layout window
2) Define parameters to control the artwork dimensions

3) Compile the macro
4) Set the parameter default values and save the macro
5) Insert the new component (macro) in a layout

6) Edit the parameters for each instance of the new component

GCC: How to build a PAM, Step 1 FYI
• Define the artwork graphically in a Layout window

• - the first step is to design your artwork macro
• This artwork will be used as an element and will be able to be

• - stretched in length
• - stretched in width
• - rotated
• - repeated

• Define parameters to effect the artwork

• Macro > Stretch… Macro > Rotate/Mirror…
Macro > Repeat... Macro > Polar…
Macro > Width… Macro > User-Defined...
• before defining parameters, draw construction lines
• use the Macro toolbar to define parameters


• Compile the macro

• - Macro>Compile
• - the macro will be saved as
<design_name>_art.ael
• - the model name is the name
used to select the artwork
from the libary

• Set the parameter default values and save the macro:

Macro > Compile
File > Design/Parameters…
In the General tab :

• define the characteristics
of the sub-network

• In the Parameter tab :

• - set the default values
• - default values are used
when the PAM is placed

GCC: How to build a PAM, Step 5&6 FYI
• Insert the new component (macro) in a layout

• Edit the parameters for that instance of the new component
• - parameters can be edited when the PAM is inserted or later

First Step : Initial Layout entry (example using Graphical Cell Compiler)

First Step : Initial Layout entry (place instance in new layout)

First Step : Initial Layout entry (assign parameters and add ports)

First Step : Initial Layout entry (enter substrate/metallization data)

First Step : Initial Layout entry (enter mesh data)

Second Step : use parametric design menu to generate designs

Only “bug” in this unsupported AEL approach is that if you are sweeping a parameter that is
unitless, you will see these errors and will need to go into each design to delete the “None” text.
THIS WILL NOT OCCUR IF YOU ARE SWEEPING A PARAMETER THAT HAS UNITS.





Third Step : prepare for Batch simulation (note that designs were
automatically added to que)

Fourth Step : Batch simulation

Fifth Step : create “discrete model” subcircuit (use DAC to access the
data)

Fifth Step : create “discrete model” subcircuit (assign a parameter to
sweep through data using File>Design Parameters)

Fifth Step : create “discrete model” subcircuit (create a symbol – optional)

Sixth Step : use “discrete model” subcircuit in top level schematic to
sweep data (use component library browser)

sweep data

sweep data (viewing swept data)

Seventh Step : use the dataset from the simulation (which interpolated
data by means of S-parameter controller) and verify results vs discrete

Seventh Step : use the dataset from the simulation (verify results vs
discrete swept data results)

Eighth Step : now you can use the interpolated approach to simulate a
new case WITHOUT using Momentum again

Eighth Step : now you can use the interpolated approach to simulate a
new case WITHOUT using Momentum again (view data)

Ninth Step : now you can even optimize the parameter (within the bounds
of the initial swept range of the parameter) of this EM-based model to
achieve a given behavior (optimizing for inductance)

achieve a given behavior (optimizing for inductance – viewing iterations)

achieve a given behavior (update optimized value and re-simulate w/o opt)

achieve a given behavior (viewing optimized results)

Tenth Step : you could also setup other parameters to be swept or
optimized to achieve other desired performance/behavior

Note: No errors encountered since these parameters have units!


Tenth Step : you could also setup other parameters to be swept or optimized to achieve
other desired performance/behavior (now use this data in a similar test bench to the
previous steps)

“That was a great capability, but is

there a more automated process?”
“Now there is! You can benefit from

both EM/circuit co-optimization and
Advanced Model Composer to
accomplish this and more!”

Momentum Optimization Another way to optimize...
Adding Layout Parameters NEW in ADS 2002C

Enables to sweep, tune or optimize geometrical variations in the
layout
- typical dimensions (length, width, gaps, spacing,…)
- interdependent layout modifications (e.g. length and width varying simultaneously)
- port locations
Two ways to create a parameterized layout component

1. Using nominal/perturbed layout artwork (Momentum Optimization)
“Nominal/Perturbed” layout parameter

2. Using existing (built-in or GCC defined) layout artwork macro’s
“Subnetwork” layout parameter


• Model Composer – ADS 2001 • Advanced Model Composer
• Fast & accurate circuit-level • Model Composer capabilities
passive parameterized models plus:
created w/ Momentum EM
• Arbitrary user-defined
technology
parameterized shapes
• Automated model generation
• Passive components including
• Patented MAPS technology spiral inductors, TFR, etc.
• Custom standard-shape • No end-user license required for
microstrip & stripline libraries use of model in schematic
and substrates
• Model generation uses AMC and
• Generates model, symbol & Momentum license
layout
• ADS Design Kit format
• Provided with Linear license


• Integrated Momentum EM Simulator • Advanced Model Composer
• Momentum RF • Arbitrary shapes
• Model Composer
• Parameterized
• Co-Simulation w/Layout Components
• Creates Design Kit
• Co-Optimization w/Layout Components
• Automated
30 GHz transition • Fast Simulation
• No end-user license required
• Momentum
• Queuing/Batch processing

• Generate libraries for wide array of custom shaped parameterized passive
models, including:
• Spiral inductors
• Thin Film Resistors
• MIM capacitors
• Arbitrary shaped
• matching networks
MomentumEM
Momentum EMengine
engine
modelingsource
modeling source
MAPSpatented
MAPS patentedtechnique
technique
forcompact
for compactmodel
modelfitting
fitting
GeneratedLibraries
Generated Librariesdon’t
don’t
requireusage
require usagelicense
license
Momentum Optimization
A Review of the Legacy Approach Flashback!
Optimization means varying physical dimensions to

achieve S-parameter goals.

A Review of the Legacy Approach: example included with ADS
C:\ADS2002\Examples\Momentum\optimization\double_folded_stub_filter_prj
A Review of the Legacy Approach: Parameters used to perturb
Start here...
1
2 Name the variable
and set the values.
Nominal layout: line.dsn

W + ∆W
W
Select the vertices that 25
20
represent the variable 4 Perturbed
Nominal (W). The result is Value
Value another layout.
Perturbed layout: line_W.dsn

A Review of the Legacy Approach: Optimization goals
Goals require Frequency values and
Response values (S-parameters).
1
Multiple goals can 2

be set up with various
weighting: For example,
W = 5 means the Error* is
multiplied by 5. The
greater the weight, the
greater the goal importance.
• Activated = 0
*Error for optimization is a • Not Activated = X
measure of how much the
goal differs from the
solution. Zero error means
the solution = goal.

A Review of the Legacy Approach: Optimization control
Optimization
status...
2
Goal achieved: Optimal value.

Finally, simulate the optimized layout
and display the results.
3
4

A Review of the Legacy Approach: Summary of Tips
GEOMETRY:
• Keep structures electrically small

• Make the drawing as simple as possible
• Take advantage of symmetries
• Make text notes in the layout, schematic, or data displays for reference
OPTIMIZATION:
• Make a written list of parameters and limit the number you use
• Perform what-if analyses in ADS circuit simulator
In general, plan the base geometry for optimization.

Create geometry that allows modification of parameters.
But there are new ways to optimize...
See Momentum co-simulation/co-optimization and Advanced Model
Composer in the main presentation. Also, see Momentum batch
simulation content in this appendix.
Agilent Direct
Americas
Asia Pacific
Australia 1 800 629 485
China 800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
Japan 0120 (421) 345
Korea 080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
Taiwan 0800 047 866
Thailand 1 800 226 008

Austria 0820 87 44 11
Belgium 32 (0) 2 404 93 40
Denmark 45 70 13 15 15
Finland 358 (0) 10 855 2100
France 0825 010 700*
*0.125 €/minute
Germany 01805 24 6333**
**0.14 €/minute
Ireland 1890 924 204
Israel 972-3-9288-504/544
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
United Kingdom 44 (0) 118 9276201

without notice.

5989-9605EN

Slides For Half Day Momentum PDF

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Slides For Half Day Momentum PDF

Încărcat de

Drepturi de autor:

Formate disponibile

Agilent EEsof EDA

Presentation on ADS Momentum

30 minutes Overview of Viewing and Using Momentum Results

60 minutes Advanced Topics [Part 1]

105 minutes Advanced Topics [Part 2]

15 minutes Final Q&A Session

Momentum Seminar momentum_01_01 Page 2

• Substrate - multiple dielectrics

• It does NOT include:

Momentum Seminar momentum_01_01 Page 3

• No simple analytical model exists

• When full 3D analysis would take too long

Momentum Seminar momentum_01_01 Page 4

Momentum Seminar momentum_01_01 Page 5

Momentum Seminar momentum_01_01 Page 6

Momentum Seminar momentum_01_01 Page 7

Momentum Seminar momentum_01_01 Page 8

Momentum Seminar momentum_01_01 Page 9

Momentum Seminar momentum_01_01 Page 10

Momentum Seminar momentum_01_01 Page 11

Momentum Seminar momentum_01_01 Page 12

Simplify hierarchical designs by

Momentum Seminar momentum_01_01 Page 13

Momentum Seminar momentum_01_01 Page 14

Europe & Middle East

Product speciﬁcations and descriptions

© Agilent Technologies, Inc. 2008

Detailed Presentation on Momentum - Part 1 of 3

• Enable regular Momentum or Momentum RF • Display Results

• Define Substrate and Metallization (pre-compute option)

Momentum Seminar momentum_01_02 Page 1

Click this submenu to toggle the analysis mode

• Solution process • Momentum MomentumRF

Momentum Seminar momentum_01_02 Page 2

Momentum Seminar momentum_01_02 Page 3

Momentum Seminar momentum_01_02 Page 4

B1(r) B2(r) B3(r)

Momentum Seminar momentum_01_02 Page 5

Method of Moments B1(r) B2(r) B3(r)

Matrix Equation ≤λ/10

Equivalent Circuit L11 R22 L22 L33

Momentum Seminar momentum_01_02 Page 7

Matrix Equation •Includes space and surface radiation

[Z] = [R] + jω[L(ω)] + 1/jω [C(ω)]-1

Momentum Seminar momentum_01_02 Page 8

1 e − jkR ≈ R1 (1 − jkR + ...)

Momentum Seminar momentum_01_02 Page 9

DC Spice Momentum RF Momentum MW

Spice model S parameters S parameters

Momentum Seminar momentum_01_02 Page 10

Momentum Seminar momentum_01_02 Page 11

Once you have created or imported your artwork…

Momentum Seminar momentum_01_02 Page 13

Student’s Guide A-36

Momentum Seminar momentum_01_02 Page 14

Conductivity defined as:

Momentum Seminar momentum_01_02 Page 15

• The formula is: We will examine a

Momentum Seminar momentum_01_02 Page 17

(Ground port component toolbar button: )

Momentum Seminar momentum_01_02 Page 18

• Single Calibrated to remove mismatch at port Edge Strip or

•Internal Not calibrated (might also call this a Edge or Strip