|
Modeling,
Analysis and Simulation Results
5.1
Coplanar Waveguide Analysis
Using Ansoft HFSS
the first stage is to draw the structure using the 3D Modeler window. Fig 5.1
shows the HFSS draw screen displaying the Coplanar Waveguide transmission line
structure.
Figure
5.1 Coplanar Waveguide
Structure in 3D modeler HFSS window
5.1.1
Drawing the geometric model
Using Ansoft HFSS
the first step is to draw the structure using the 3D modeler window Fig. 5.1
shows the HFSS draw screen displaying the Coplanar Waveguide TL Structure.
Before drawing the model solution type need to be specified. There are three
solution types in HFSS version 9, Driven modal, Driven Terminal and Eigenmode.
For this particular structure we have used driven modal solution type. The
second step is to specify the units of measurements for drawing geometric
models. The units used here for this structure are millimeters (mm).
After inserting an
HFSS design into the 3D modeler window the first step is to build the modal as
a collection of 3D objects. Each material type is treated as a separate object.
For this structure we have drawn a box by selecting two diagonally opposite
corners of the base rectangle and then specify the height, immediately after
this the properties dialog box appears automatically, enabling us to modify the
object’s properties. Two boxes of 0.6mm thickness and 5.0mm width were drawn
and named as ground planes; the central conductor is a box of 0.6mm thickness
and 100mm width as shown in figure 5.1. Similarly the dielectric substrate of
0.56mm thickness were drawn beneath the conductor and ground planes and
assigned Gallium Arsenide material properties. Lastly drawn an outer box, which
is 4-times in height above the conductor and ground planes and 2-times deep
below the substrate as shown in the figure 5.1.
5.1.2
Defining Materials
The material of
which the structure is made has to be specified. When you assign a material to
an object, you can specify whether to generate a field solution inside the
object or on the surface of the object. If you select to generate a solution
inside the object, HFSS will create a mesh inside the object and generate a
solution from the mesh. If you select to generate a solution on the surface of
the object, HFSS will create only a surface mesh for the object. By default
solve inside is selected for all objects with a bulk conductivity less than
siemens/meter and for perfect insulators.
Every object by
default when drawn is assign material property of vacume. We can change then
the desired material. For this structure we have assigned a gold material to
the central conductor and to the ground planes. A GaAs material was assigned to
the substrate and the outer box was set as an air box.
5.1.3
Defining Ports and Boundaries.
The
boundaries for the structure must next be assigned and also the ports
calibrated and defined. Figure 5.1 displays the location of various ports and
boundaries for the Coplanar Waveguide structure. The structure has two ports,
port 1 is for the signal to enter and port 2 is for the signal to exit the
structure. Port 1 and port 2 are both 50W.
Boundaries for the surface of the structure are defined, to let the simulator
take into account the characteristics of the materials of which the structure
is composed. Note that the structure under consideration has an inner and outer
metal boundary (Perfect E) and is internally composed of air.
There are two
recommendations for assigning waveports to the structure. First is the wave
port size should not exceed lambda/2 in any dimension, to avoid permitting a
rectangular waveguide modal excitation, 3(2g+w). As shown below in Fig. 5.2.
Secondly the
waveport width should be no less than 3´
the overall CPW width, or 3´(2g+w),
and port height should be no less than 4´
the dielectric height, or 4h.
For port field
display we need to assign integration line when defining waveport. An
integration line is a vector that can represent a calibration line that
specifies the direction of the excitation field patterns at a port. Also
for the computation of Zpi and Zvi port impedances, we need integration line.
[80]
5.1.4
Background objects and boundaries
The system creates
the background objects itself and is assigned the material characteristics of a
perfect conductor. It is the region that surrounds the geometric model and
fills any space not occupied by an object. There are many types of boundaries,
e.g. perfect E, perfect H, finite conductivity, impedance, lumped RLC etc, but
in this report we have used perfect E, and radiation boundaries as well where
appropriate.
5.1.5
Modal field distribution
CPW E-field
distributions are shown in the figure below, Fig.5.3.
Figure
5.3 E-field
distributions
at port1 for CPW
·
Basic vector fields can be displayed in setup
executive parameter/port impedances, by highlighting a port and using the port
field’s button beneath the graphical window to adjust the window.
·
E-fields in CPW are symmetric to either side
of the center trace, with E fields extending from each ground to the trace in
phase. H-fields circulate around the center trace.
In figure 5.3 all
the arrows in the plots are of different sizes and the length of the arrows
vary. This simply indicates the direction and field’s strength respectively.
5.1.6
Defining Mesh Operation
Mesh operations
are optional mesh refinement setting in HFSS that provide HFSS with mesh
construction guidance. This technique of guiding HFSS’s mesh construction is
referred to as “seeding” the mesh. Seeding is performed using the mesh
operations commands on the HFSS menu.
5.1.7
Specifying Solution Setup
HFSS computes a
solution by adding a solution setup to the design. Each solution setup includes
the general data information about the solution’s generation; adaptive mesh
refinements parameters and frequency sweep parameters. Then a solution sweep
can be added to the solution setup.
5.1.8
Setting up 3D Simulation
The Fast Frequency
Sweep is preferred for the analysis of structures. This is because it is much
faster to solve than the discreet frequency sweep. The fast frequency sweep
saves time by solving problem at a single frequency. This is achieved by a
rational function approximation for the frequency bandwidth of interest. The
fast frequency sweep only calculates other frequencies in the specified
bandwidth approximation. The software interpolates the sweep frequency points
and uses additional frequencies to get a high degree of accuracy and other
frequencies other than the single frequency specified.
In the simulations
performed in this report, the fast frequency sweep was employed and all
problems solved for the dominant or first order mode. All field solutions were
saved for each analysis, which were then processed using the HFSS
post-processing facility.
The following
set-up conditions were used for CPW structure.
Start frequency = 1GHz
Stop frequency = 35 GHz
With
0.1 GHz step size and 0.01 maximum delta S per pass.
The initial mesh
generated by Ansoft HFSS, solved by the finite element method, which was
discussed earlier, was used to solve the structure three-dimensionally for each
different value of dielectric constant. Using a fast frequency sweep from 1GHz
to 35GHz and ensuring the port impedance was 50W,
both for port 1 and port 2 the transmission line structure were analyzed. The
simulation results for CPW are shown in the figure 5.5 below.
Figure
5.5
Magnitude S-parameter graph of CPW.
The red and green
lines represent S12 and S21 respectively. While the black and blue lines
represent S11 and S22 respectively. The S12 and S21 show the transmission
behavior and S11, S22 predicts the reflection behavior at port1 and port2
respectively. The simulated results are in good
agreement with the measured results for CPW.
5.2
Analysis of EBG Structure with two-rows of holes on each ground plane
We used the CPW
structure and etched two rows of six holes on each ground plane and simulate
the structure with the help of High Frequency Structure Simulator (HFSS)
software. The simulated results are very similar to the measured results. The
EBG structure with two rows of six holes on each ground plane is depicted in
the figure (5.6) below.
Figure
5.6 EBG structure with two rows of holes on each ground plane.
The dimensions and
materials, of the substrate, central conductor, ground planes and boundaries
are given in the table (5.1) below.
|
|
Air
|
|
Substrate
|
Gallium Arsenide (GaAs)
|
|
Ground Plane 1
|
Gold
|
|
Ground Plane 2
|
Gold
|
|
Central Conductor
|
Gold
|
|
Cylinder Radius
|
|
|
Cylinder Height
|
0.5mm
|
|
Substrate Thickness
|
500mm
|
|
Central Conductor
& Ground plane Thickness
|
0.5mm
|
|
Outer Boundary
|
Perfect Conductor
(Perfect E)
|
|
Number of Wave Ports
(Excitations)
|
Two
|
Table
5.1 EBG
structure with two rows of holes measurements.
The simulated
results are compared with measured results and both the results resembles very
closely. For every structure the measured results are presented first followed
by the simulated results. For the structure with two rows of holes on each
ground plane the measured results are given in figure (5.7) followed by the
simulated results in figure (5.8).

Figure 5.7
Measured results for the EBG structure with two rows of holes on each
ground plane.
The red and green
lines represent S11 and S22 respectively, while the blue and pink lines
represent S12 and S21 respectively. In the S-parameter plot we can see that
there is a smooth transmission of signal before 14.2GHz and the transmission
line shows a filter behavior after this point. The reflection coefficients are
increasing slightly with increasing frequency and exactly at 15GHz are constant
near 0dB.

Figure
5.8 Simulated S-parameter results for EBG structure with two rows
of holes on each ground plane.
Here the black and
blue lines represent S11 and S22 respectively while red and green lines
represents S12 and S21 respectively. The simulated results shows nearly the
same behavior as the measured results, the peak above 0db in simulated results
is due to the perfect E boundary used as a boundary for the structure
simulations in HFSS. The higher order propagation modes are reflecting back due
to the perfect conductor boundary and cause the peak above 0db, if we use
radiation boundary instead of perfect E boundary we can eliminate this peak.
For this structure there is smooth propagation of signal until14GHz and the
transmission line is showing filter behavior after this point. The reflection
and transmission behaviors are nearly the same as in measured results.
5.3 Analysis of EBG Structure with three-rows of holes on each ground plane
Again we used the
CPW structure and produce three rows of holes on each ground plane and simulate
the structure with the help of HFSS software. The simulated results are very
similar to the measured results for this structure as well. The reflection and
transmission behaviors in simulated plot are very similar to the measured
results as shown in figures (5.10) and (5.11) respectively
Figure
5.9 EBG structure with three rows of holes on each ground plane.
The dimensions and
materials of the substrate, central conductor, ground plane and boundaries are
given in the table (5.2) below.
|
Outer Box
|
Air
|
|
Substrate
|
Gallium Arsenide (GaAs)
|
|
Ground Plane 1
|
Gold
|
|
Ground Plane 2
|
Gold
|
|
Central Conductor
|
Gold
|
|
Cylinder Radius
|
500mm
|
|
Cylinder Height
|
0.5mm
|
|
Substrate Thickness
|
500mm
|
|
Central Conductor
& Ground plane Thickness
|
0.5mm
|
|
Outer Boundary
|
Perfect Conductor
(Perfect E)
|
|
Number of Wave Ports
(Excitations)
|
Two
|
Table
5.2
EBG structure with three rows of holes measurements.
The
same procedure is adopted for this structure as well, the simulated results are
presented after the measured results for comparison and for this case the
results are again showing very similar behavior.

Figure
5.10 Measured S-parameter results for the EBG structure with three rows of
holes on each ground plane.
Figure
(5.10) shows the measured results for the structure with three rows of holes on
each ground plane and here again the transmission line is showing a filter
behavior at 14.2GHz. The reflection coefficients S11 and S22 are nearly
constant after 15.2GHz and showing frequency blocking behavior.

Figure
5.11 Simulated S-parameter results for EBG structure with three rows of
holes on each ground plane.
Again for this
case of three rows of holes on each ground plane the simulated results shows
nearly the same behavior as the measured results, the peak above 0db in
simulated results is again due to the perfect E boundary used for structure
simulation. The higher order propagating modes are reflecting back due to the
perfect conductor boundary and cause the peak above 0db, if we use radiation
boundary instead of perfect E boundary we can eliminate this peak. In the plot
above we can see that the propagation parameters (S12 and S21) are constant
until 14GHz and showing filter behavior afterwards. The reflection parameters
(S11 and S22) are 0dB at 14.4GHz to 19GHz and showing frequency blocking
behavior in this region.
5.4 Analysis of Vertically Periodic Defected Ground Structure
(VPDGS)
The prominent
feature of the proposed structure is that it is possible to organize the
periodicity along the vertical direction as well as the horizontal direction.
It is named as Vertically Periodic Defected Ground Structure (VPDGS). On the
other hand the structures used in this report earlier, EBG structures have only
horizontally periodic holes, i.e. serially cascading structure along the
transmission direction [81]. Here we have considered the simplest structure
first, with one hole on each ground plane and then moved on to the structure
with two holes vertically on each ground plane. More complex structures with
two rows of holes vertically and horizontally and three rows of holes
vertically and horizontally are given in appendix1. The simulated results are
in good agreement with the measured results and shows the resonator and filter
behaviors for some particular frequencies. Then we changed the distances of
holes with respect to each other and noted some interesting results.
5.4.1 VPDGS with one hole on each ground plane
The
structure shown below is the simplest structure with only one hole etched on
each ground plane. The structure is symmetric along both the axis. The
substrate used is Gas
(
and H=0.64mm) and the central conductor
and ground planes have assigned the gold material properties. The same setup
was used as for the earlier EBG structures. The measured results are shown in
figure 5.13 followed by the simulated results in figure 5.14.

Figure
5.12 VPDG Structure with one hole on each ground plane.

Figure
5.13 Measured results for the VPDGS above.
Figure
(5.13) shows the measured results for the structure shown above.
| The periodic
pass and stop bands, and steep cutoff rejections are observed in the
performances, which are in good agreement with each other.
|
Figure
5.14 Simulated results for the
VPDGS Structure with one hole on
each
ground plane.
The
red and green lines represent S12 and S21 respectively while black and blue
lines represents S11 and S22 respectively. The resonator behavior in the
measured results occur at 4GHz while in simulated results it is occurring at
4.4GHz, this is due to the differences in the setup and software used for the
simulations.
Now
the distance between the central conductor and hole is reduced from 1.5mm to
0.5mm for both the ground planes and simulate the structure using the same
setup and boundary conditions. We noted for this new structure that, with
changing the distance between the central strip and holes, the resonator peak
moved from 4.4GHz to 5.9GHz. So by changing carefully the distance among the
holes, we can design microwave circuits like resonator, filters etc using EBG
structures. The structure with reduced distance between central strip and holes
is shown in the figure below.
Figure
5.15
VPDGS Structure with hole distance reduced to 0.5mm.
Figure
5.16 Simulation
results for the structure with reduced (0.5mm) distance.
The
red and green lines represent S12 and S21 respectively while black and blue
lines represents S11 and S22 respectively. Changing the distance between the
central strip and hole by 1mm causing the resonator peak to displace by 1.5GHz.
5.4.2
VPDGS with two holes on each ground plane
After analyzing
the structure with one hole on each ground plane now we will analyze the case
of two holes vertically on each ground plane. Again for this structure the same
setup and boundary conditions are used. The measured results are followed by
the simulated results and shows very similar behaviors.

Figure
5.17 VPDGS structure with vertically two holes on each ground plane.
Figure
5.18 Measured
results for the structure with vertically two holes on each ground plane.
Figure
5.19 Simulated results for the structure with vertically two holes on each
ground plane.
The
black and blue lines represent S11 and S22 respectively while red and green
lines represents S12 and S21 respectively. There are very small differences
between the measured and simulated results. The propagation parameters of
simulated results are 0.5GHz displaced towards the highest frequency. Now
analyzing the effect of changing the distance between central strip and holes.
The distance is reduced from 1.5mm to 0.5mm as shown in the figure (5.20)
below.
Figure
5.20 VPDGS
structure with holes distance reduced to 0.5mm.

Figure
5.21 Simulation results for the structure with hole distance
reduced
to 0.5mm
The
red and green lines represent S12 and S21 respectively while the black and blue
lines represent S11 and S22 respectively. By reducing the distance between the
central strip and holes by 1mm the S12 and S21 parameters are displaced by
0.9GHz towards the highest frequency. Similarly S11 and S22 are moved to 6.6GHz
from 5.7GHz. So changing the hole’s distance with respect to the central
strip we can design very interesting circuits for microwave and millimeter wave
applications.
5.5 Summary
CPW structure is
used as a reference modal and simulated with the help of HFSS software.
Electric field distribution and modeling procedure are described for CPW
structure and the same modeling steps are used to simulate the EBG structures.
Firstly EBG structure with two rows of holes on each ground plane is analyzed
and then structure with three rows of holes on each ground plane is considered.
The measured and simulated results for both the structures are very similar.
The VPDGS are simulated and compared with measured results. Also the effect of
changing distance among the holes and central conductor are analyzed. In
conclusion, the measured results shows that these new EBG transmission lines
are 30 times better than the conventional coplanar waveguides and we can design
different active devices by etching holes in the ground planes which can reduce
space on the printed circuit board.
|