Jump to content

Johan Huysamen

Moderators
  • Content Count

    68
  • Joined

  • Last visited

  • Days Won

    3

Johan Huysamen last won the day on August 4 2016

Johan Huysamen had the most liked content!

6 Followers

About Johan Huysamen

  • Rank
    Advanced User

Profile Information

  • Gender
    Male
  • Country
    South Africa
  • Are you University user?
    No

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Dear RF Engineer In your Feko model you are using the "Planar Multilayer Substrate" option to model an infinite substrate and ground plane. Could you please confirm that your CST results where also obtained using an infinite substrate/ground implementation? Kind regards, Johan H
  2. Hi celaleddin taylan From your three images it looks like you have changed the frequency ranges, but kept the number of frequency samples the same. If you then reduce the range, you sample the frequency more finely. The behavior you see is caused by the fact that the number of discrete frequency samples is insufficient to locate the minimum reflection coefficient. More samples (finer sampling) is required. I would suggest that you try the "Continuous (interpolated) range" option for the solution frequency: With this option Feko will automatically sample a sufficient number ot frequency points to accurately locate the minimum value (at the resonance point).
  3. Hi Kuzi From your description it sounds like you are trying to feed many antenna elements in an array. Please note that the SParameter configuration is specifically for requesting and calculating S-parameters between multiple ports. It is not intended for exciting antenna elements. This is done using a StandardConfiguration. For your application you should consider using the Finite Arrays tool: This allows you to create a single antenna element and then specific the number of array elements as well as the distribution (including the magnitude and phase of the element excitations). More information on this tool can be found under the Feko Help. You could also have a look at Example A17 of the Feko Example Guide. Kind regards, Johan H
  4. Dear sutton304803 Various options for infinite planes and (dielectric) half spaces can be selected through the "Plane / ground" option under the Construct tab in CADFEKO: This opens the "Plane / ground" dialog where you have various options. For your application you could use one of the two "Homogeneous half space in region Z<0" options or even the "Planar multilayer substrate" option if it is specifically required to have the top of the half space at any other value than Z=0. More information on these option can be found under the section "Infinite Planes and Half-Spaces" in the Feko Help (accessed by hitting <F1> in CADFEKO).
  5. Hi ynathan CADFEKO has an option to import one CADFEKO model into another. On the Home tab select Import -> CADFEKO model (*.cfx)... This will allow you to import various entities from one CFX file into your current model: There are also options to merge identical variables and media to ensure that you not have duplicate identical denifition as well as an option to add a prefix (text) to all the imported item labels to distinguish them from the original model items. Kind regards, Johan H
  6. Hi Patrick Theoretically a plane wave comes from infinity and is incident onto the geometry. The plane wave is applied as an excitation onto the geometry (mesh) in Feko. As such, it is applied to, and interacts with, the mesh where-ever there is a mesh element in the model. Similarly, when a (near) field result is requested, the plane wave will be considered in the calculation of that near field point where-ever required. Consider a Perfectly Electrically Conductive (PEC) sphere with radius r meshed into surface mesh triangles and solved with the default Method of Moments (MoM) in Feko. If we have an incident plane wave as excitation and we sample near fields say on a spherical surface at a radius of 2*r, the plane wave is applied as follows: Mathematically the plane wave excitation is evaluated at the positions of the mesh triangles and applied to the MoM solution matrix as an excitation (known as the right-hand side vector) The MoM is solved to compute the induced surface currents on the mesh triangles. The near fields are calculated by considering two contributions at each required near field request point: the (re)radiation of the induced surface currants and the contribution of the incident plane wave In effect then the plane wave does not have a physical extent. It is applied and evaluated by Feko where-ever it is required based on the mesh and requested results. Kind regards, Johan H
  7. Hi, You should set the thickness on each face to be the actual physical thickness of the conductor (and not half on each). Feko will model an equivalent surface current modified to account for the physical thickness as specified. Kind regards, Johan H
  8. Dear tmarinovic With your example I believe you are triggering: WARNING 40147: Near fields for spherical modes are computed in the cut-off region beta*R < N, reduce spherical wave order N or increase distance R Note that your OUT file should include details on the value for R for your specific example. In Feko the spherical mode source is intended as an option to represent an antenna installed on a large platform for efficient simulations of the installed performance of such an antenna. Here the focus is specifically on efficient far field calculations and not on near field sampling relatively close to the antenna. Details on Feko's implementation for spherical mode sources can be found under the "AS card" section starting on p986 of the Altair Feko User Manual. (UserManual.pdf located in the ..\help\feko\pdf folder of your Altair installation.) Unfortunately we are not able to go into more details regarding Feko's implementation - other than sharing this (publicly available) information. Kind regards, Johan H
  9. Dear tmarinovic From the fact that you are referring to a complex quantity for the near field, I assume you are comparing only a single component. Could you please confirm which component this is (E_r, E_theta or E_phi)? Also, could you confirm the unit of the values you have mentioned? Are these in V/m? In general, sampling near field calculations too close to a spherical mode source is not recommended. Kind regards, Johan H
  10. Hi Bidisha From your description I assume that you would like to extract the multi-port Z-parameters. (Note that the "MoM Matrix" solved during the Method of Moments solution in Feko is also known as the "Impedance matrix" as the elements of this matrix have the the same units and impedance.) To extract the Z-parameters you should request S-parameters in CADFEKO and then, after the simulation has completed, in POSTFEKO use the Export -> Touchstone option. Under the Touchstone export dialog there is an option to select the "Network parameter type" for the export. There you should select "Z (impedance parameters)" to then export the data to a Touchstone file. Kind regards, Johan H
  11. Hi zml In addition to Torben's comments above, it is recommended to not have the Edge port electrically too wide. I would suggest implementation the feed across a thinner edge as in the following image. You will also notice that the feed sections are lifted off the substrate. While this should improve the Feko model, even this does not appear to replicate the results from the paper. I suspect that the HFSS simulation used some modal excitation that impresses the required microstrip mode. As an equivalent option is not available in Feko, the best approach would be to try to replicate the measured result by including an SMA connector in your Feko model. This could then be excited with a (coaxial) Waveguide port or a (Finite Element Method) FEM Modal port. Some example are shown below. You might have to contact the authors for information on how the SMA connector was physically connected to the antenna for the measurement. If you replicate exactly that in Feko, we would expect a good agreement between our Feko model and the measurement.
  12. Hi Theodoros You mention that you have an infinite cylinder inside a circular loop. If this is modeled with the PBC in Feko it should be noted that the loops will also be repeated by the PBC. Your two models would then be different as one includes a loop repeated every h1 and the other has a loop repeated every h2 along the length of the infinite cylinder. Unless your initial model does not include the loop(s) - only having the cylinder? In that case you might see small differences due to different meshes being used. If the above does not clarify the issue, please share your model if possible. Kind regards, Johan H
  13. Hi bouvy The recommended approach would be setting the internal regions to "Free space" and the surface faces to your "Copper" metallic medium with a specific thickness. If the structure is intended to be solid metal then, yes, you should use a thickness larger than the penetration depth. I suspect that you are receiving WARNING 2757 because you have some small triangles in your mesh. If the size of the mesh triangle is small compared to the thickness defined for the metallic medium, then this warning is issued. I would suggest that you check your mesh settings (including the curvature refinement settings under the Advanced tab of the "Create mesh" dialog to ensure that your mesh does not include very small mesh element. Additionally I would suggest that you use thickness that represent the actual structure where possible. For your vehicle model I would assume that the structure is actually hollow with a wall thickness of a few centimeters or millimeters. Note that you could also define these thicknesses differently per individual face - using actual thicknesses for shells that you now the wall thickness of and values larger than the penetration depth for solid parts. Kind regards, Johan H
  14. Hi RF Engineer This ERROR 40161 is a general error and usually indicates that the model includes some setup item or setting that is not supported along with Characteristic Mode Analysis. Are you able to share this model? Kind regards, Johan H
  15. Hi ferdaousAbderrazakLajili I noticed two issues with your FEKO model: There are small gaps between your ground and antenna trace faces and the substrate. You should create these faces to lie exactly on the substrate surfaces You have the frequency range set as being from 0.4 to 10. Note that these values are in Hz. Maybe you want 0.4 to 10 GHz and should therefore enter 400e6 and 10e9 respectively. In the paper - as seen in Fig. 3 - 100MHz to 2 GHz (100e6 and 2e9) was used. For creating a coaxial excitation you have various options on FEKO. For the most accurate representation I would suggest physically building a short coaxial section (inner and outer conductor cylinders with a dielectric region in between. Using the default Surface Equivalence Principle (SEP) this can then be excited either by a Waveguide Port on the back coaxial face or by an Edge Port located between the inner and outer conductors. (Note that the Waveguide Port might not be supported with the Characteristic Mode Analysis in FEKO). For examples on these and other ways of feeding antennas, take a look at the FEKO Examples Guide (ExampleGuide.pdf located in the \help\feko\pdf folder of your Altair installation).
×
×
  • Create New...