Introducing Antenna Magus Presenter Location Date
Overview What is Antenna Magus? The design problem An Antenna Magus Demo Find Design Export Arrays, tools and Adding your own antenna Highlighting some recent extensions Ensuring quality of models and designs A closer look at some synthesis approaches
What is Antenna Magus? Antenna Magus is the first antenna design tool of its kind. Antenna Magus allows antenna engineers to find + learn about + design many antennas, and export models of designed antennas to EM simulation tools like FEKO Engineers may also add their own antennas to the database
The design problem Antenna Design The process of creating an arrangement to achieve a desired effect Antenna Analysis Take a given arrangement and predict its effect Frequency Gain Lengths Angles Antenna Magus aids in antenna design Lengths Angles Frequency Gain Analysis is part of the design process
The faces of Antenna Magus Find Info Design Performance Export Array Synthesis Libraries Tools Add your Own
Demo Find Design Learn Export
Prototype
Stages in array design Layout assuming isotropic elements Replace isotropic elements with real element pattern (in isolation) Calculate mutual coupling between elements Compensate element patterns for coupling effects Feed network design etc.
Example Choose a concentric circular layout, with main bean steering and side lobe control
Layout, distribution and Isotropic pattern result
Choose microstrip patch as element
Synthesized pattern This array can now either be taken to the next design step (mutual coupling effects) Or its pattern exported as a source (or load) in another simulation
Toolbox Friis equation Chart tracing Radar equation Gain/BW Pattern calculator Aperture distribution Gain from an aperture
Toolbox: Example Gain/BW
Custom antennas in Antenna Magus Use your antenna in Magus Information and documents Designs FEKO models Performance data
Custom antennas in Antenna Magus Document and store antenna designs Collaborate and share information and models Work with your antennas inside the Antenna Magus workflow
Recent extensions Version 1.0-68 Version 2.0-113 Version 3.0-148 Version 4.0-200
Recent extensions Pre-optimised designs for specific applications (e.g. WLAN) Instant performance estimation for certain designs Any general 3D radiation pattern may be used to represent the elements in an array synthesis (various file formats supported). The total number of antennas in each search group is indicated in the find mode.
Recent extensions IEEE Axial ratio in db (handedness not included) can be plotted 2D and 3D Plots of co- and cross-polarised gain based on the Ludwig III method. Many additional export options 3D Pattern files -> VSS format, CSV format, IEEE1979 format Array layouts -> XML format Additional sampling options when exporting 2D data
Ensuring quality of Models and Designs Simulation models are first tested against published results Validation criteria are set e.g. S11 < -15dB, fc < +- 5% etc. Validation sets are created (between 50-250 per group depending on the number of objectives) and simulated. Generate models Compare results to expectation Sample designs Run simulations
Validating designs and simulation models Result tables are created that are scrutinised by the engineer Failures are investigated modelling problem design problem. If the same failures occur for different solvers or techniques it usually points to a design error or invalid constraint. Failures unique to a technique usually points to a model error or shortcoming Over the whole design range: The designs algorithms work The models are correctly parameterised The meshing is correct The best techniques are used The basic output requests are correct Different solvers and techniques agree Etc
Synthesis of parabolic reflector antennas A basic reflector: The focus-fed parabolic
The focus-fed parabolic antenna Let s take look at one design with these inputs: Frequency Gain (G) Sidelobe level (SLL) Feed beamwidth (FBW) Edge taper (ET) Feed distribution efficiency (FDE)
The focus-fed parabolic antenna The design process G FBW Gain = SLL 39.7 (40) ET dbi FDE Consider aperture efficiencies. F/D ADE Consider aperture distribution shape. SLL = -20.05 (-20) P db BR Blockage ratio is calculated and used to determine and compensate for overall efficiency. D F EFF The design assumes an ideal pattern-excitation (i.e. no horn) but feed-blockage is compensated for and can be adjusted in the analysis. Real horn-feed Blockage adjusted
Choosing design inputs and outputs There are many feasible combinations of inputs that could be the basis for a design! A careful choice needs to be made of the most useful combinations Design approaches for each combination allow for flexible usage in practical situations
Synthesis of parabolic reflector antennas A complex dual reflector: The Cassegrain
The Cassegrain reflector Design for gain (38 dbi) at a specific frequency There are many viable designs to achieve the required performance. The choice between these designs rests on external factors and implications. flexible input options allow case-specific factors to be considered.
Synthesis with simple dependencies Consider a simple pin fed rectangular patch Closed form solution based on transmission line theory and simple slot current model for radiation resistances. Model is not 100% accurate, but works well enough to provide a first order design for a wide range of inputs.
An antenna with simple dependencies Strong relationship between pin location and input impedance allows adjustment of the pin inset to adjust the impedance Only a moderate effect on other performance properties like resonant frequency. The design approach is robust and works! Frequency -> patch dimensions Impedance -> pin position Etc.
An antenna with complex dependencies Consider the aperture coupled patch This patch has several independent parameters that all have dependent effects. Some first order effects of modifying a parameter are well known, BUT secondary effects are considerable. E.g. Increasing aperture size reduces the radiating resistance AND the resonant frequency! It is extremely complex to create a robust algorithmic design for this antenna!
Synthesis with complex dependencies Circuit models separate the problem into separate components: microstrip line + slot + patch ; Each have complicated design equations, or require iterative optimisation to resolve. Quantities must be derived from physical parameters. E.g. simple slot current model used for the pin fed patch falls apart with the much thicker substrates that are used in the aperture coupled patch. Derivation of the coupling from the feed-line to the patch through the slot, is formidably complex! Empirical and analytic relations between circuit model quantities and physical parameters only applicable to very specific combinations of material parameters. E.g. for er ranges of 2.1-2.3. Not good enough to base a general design on! D. M. Pozar, Microstrip antenna aperture-coupled to a microstripline, Electronics Letters, v21 n2, 1985 pp. 49 50. M. Himdi et al., Analysis of aperture-coupled microstrip antenna using cavity model, Electronics Letters, v25, n6, 1989, pp. 215 216
Multidimensional regression: Some drawbacks
Input 2 Radial basis function (RBF) regression 7 Required value 3 8 Input 1 1 Known designs Design space Weighting values based on distances to known designs 7 0.07 3 0.4 8 0.53 1 0.1
A practical application of RBF s Step 1: Determine limits of input values that can be designed for. Step 2: Choose a sparse set of design points in the chosen space. Complete a satisfactory design for each of those points. Step 3: Choose a random design point inside the design space. Step 4: Apply the radial basis function interpolation using all of the available designs. Evaluate this design. Step 5:If the design does not meet specification, adjust the design till it is satisfactory and add it to the design set. Select bounded design space Known/tested designs Repeat step 3-5 until the interpolation always yields satisfactory designs During the design testing process the design space can continually be adjusted Great emphasis must be placed on the accuracy of the computational models used to analyse design points. What about frequency? As an added dimension it increases the sample space. Normalizing by frequency could lead to Non-unique bestdesigns solutions for non-frequency-scalable structures. You need a good validation/testing system to coordinate the simulations and designs! Add to known designs Improve design Test RBFbased design quality Choose random design point New designs are acceptable
A practical application of RBF s The aperture coupled patch requires 40 or 50 known design points in a large 5D design space! Input quantity Minimum Maximum Operating frequency 500 MHz 20 GHz Top substrate relative permittivity 1 4.4 Top substrate thickness 0.15 mm 90 mm Bottom substrate relative permittivity 1.8 13 Bottom substrate thickness 0.127 mm 24 mm
Thank you More information: www.feko.info/antennamagus Contact information for local distributors: (FEKO distribution network) http://www.feko.info/about-us/contacts