|
The 2000 CAD Benchmark Unveiled
Over the past few years the series of CAD Benchmarks in Microwave
Engineering Europe has proved an enduring favourite among our readers,
none more so than those challenging the capabilities of 3D electromagnetic
simulators. This year for the first time we have chosen to benchmark
the performance of these packages in simulating a free
space radiation
problem - a balanced antipodal Vivaldi antenna - whose characteristics
are described below. This problem has been circulated to a broad
group of CAD vendors, and we will publish their solutions in the
November issue. Readers are invited to attempt their own solutions
to the Benchmark, and can post solutions, comments or questions
to the vendors on our Web site at www.mwee.com.
Send us your comments on this article
Download the article in Adobe PDF
format
[approx 175 kB]
 |
|
Figure 1. 3D representation
of balanced Vivaldi antenna. Yellow is stripline track/flare,
brown is lower groundplane transition/flare, and purple (unshaded)
is upper ground plane transition/flare, which is the same
as the lower groundplane. Cyan is the substrate outline.
|
The conventional Vivaldi antenna typically has at least an octave
bandwidth making it suitable for a wide variety of applications,
but can be limited by its radiating slot transition. The antipodal
Vivaldi, in which a microstrip line and its groundplane both gradually
flare out, removes the bandwidth limitations of the transition.
The lower frequency
limit is now determined by the cut-off mechanism
of the flare, namely that the aperture is half a wave-length wide.
However the skew in the electric fields across the slot leads to
poor cross-polar performance, which also degrades as the frequency
rises.
A development of the antipodal Vivaldi which remedies its worst
characteristics is the subject of this Benchmark. Converting it
to a triplate based
structure, by adding in an additional dielectric
and metallization layer which balances the E-field distribution
in the flared slot, greatly improves its cross-polar performance.
The antenna starts in stripline. One side of a board has the input
track which is then flared to produce one half of a conventional
Vivaldi. On the other side (and on the second substrate) the ground
planes are reduced down
to form what will be a balanced set of lines.
The 'ground lines' are then flared out in the opposite direction
to form the overall balanced antipodal Vivaldi.
The configuration uses arcs and elliptically tapered geometry which
can be a good test for some solver software and, along with the
aim of predicting certain far-field characteristics, led us to conclude
that this would be a non-trivial but feasible benchmark.
Another
interesting aspect will be to see how the Vendors determine their
air-boxes, radiation boundaries etc, unlike a closed waveguide or
stripline problem. This is usually an arbitrary aspect in most software
and a source of uncertainty and error for inexperienced users.
Dimensional data
Figures 2 to 5 give most of the dimensional detail. The key points
are as follows:-
- Substrate is 40mm wide. A length of 90mm is adequate for defining
all the detail
- Stripline track width is 3mm
- Output flares are elliptical arcs 30mm long by 9mm high, representing
a 3.33:1 major/minor axis ratio
- Transitions from groundplane to balanced line are also elliptical
- Aperture slot width is 15mm
- Port/edge info: apart from the input edge, all other edges are
exposed dielectric. At the input edge metal is present connecting
the two ground planes together except for a 12mm wide area centred
around the track where the excitation port is present (a SMA edge
launcher in practice). See Figure 4 for this detail.
 |
|
Figure 2. General geometry and aperture
detail.
|
Figure 3. Central (stripline) metallization
layer. Note: Output flare and radius also to the
groundplane layers.
|
Substrate
The substrate used in this example is Rogers Duroid 5870:
 r
=2.32, width 40mm, length 90mm. Two 1.575mm (0.062") thick boards
are used to create the overall triplate structure. Metallization
thickness and
loss tangents may be ignored if necessary. Other
than the edges near the port mentioned above, no other edge
metallization or vias are present. Examples of these antennas
have been fabricated on higher
r
substrates as well, particularly in array applications.
 |
|
Figure 4. Outer ground plane layers.
Note: There is a short straight section between the end
of the transition and the start of the aperture flares.
|
Figure 5. Perspective view and input
port edge detail. Note: Stripline track layer is sandwiched
by two 1.575mm thick duroid layers.
|
Outputs
A reasonable amount of background information is requested
to
make the assessment more informed, as well as the analysis
results themselves. The following outputs are expected from
the CAD vendors:-
- Model geometry printout with annotation which would confirm
dimensions/ approximations used, boundary regions/definitions
(e.g. air volumes, excitations, radiation boundaries etc), as
well as feedback on ease of definition etc.
- S-Parameters and input
impedance: A minimum analysis range of
0.5 -10GHz is required, but optionally higher (the original device
was characterised up to 18/20 GHz). Plots (in dB vs. frequency
format), and Touchstone format files for return loss are needed
(to make replotting and comparisons easier)
- Pattern data output at 10.0GHz for E- and H-plane cuts, co-
and cross-polar, gain, beamwidth. plots should be in Cartesian
format in dB vs. degrees. Optionally 3D colourful plots/animations
may be submitted electronically to illustrate features. As this
is principally an antenna benchmark, emphasis will be on these
results.
- Comments on problem entry/setup, solution process etc, discussion
of results
- Hardware data: make/model, operating system, clock speed, RAM
fitted, number of CPUs
- Software data: name, version, elapsed run times, RAM/disk space
needed, number of elements used (or cells/ unknowns)
The first part of this Focus can be viewed
here
.
« October Issue
|
