Jump to content


  • Content Count

  • Joined

  • Last visited

  • Days Won


mel last won the day on April 3

mel had the most liked content!

1 Follower

About mel

  • Rank
    Super User

Profile Information

  • Gender
  • Country
    South Africa
  • Are you University user?

Recent Profile Visitors

1876 profile views
  1. Please rerun the model capturing the near fields with the correct port settings and then verify if the original problem is still there. If so, please attach the updated files.
  2. This error could happen if your model is on a network drive.
  3. 1. Why do you want to use a solid metal? The recommended method is a surface definition similar to Example Guide model A-02 in the examples folder. A solid model attempting to model a lossy metal (very high conductivity) would require extremely small mesh elements to correctly capture the rapidly attenuating fields. This would drive up the computational resources. 2. Have a look at the aperture coupled patch on finite dielectric, Examples Guide model A-10.
  4. Yes, write the m and n values. Whether it is TE or TM is done automatically. The higher m and n, the smaller the mesh size required will be.
  5. It may not be related to the original problem, but in the last model you attached, why are there two apparently identical sources on the same port? If you meant to place wgsource2 on the opposite port, where is the energy supposed to go with both ports active? Or did you mean to make the bottom port absorbing? You don't need to place a source there then if it is only meant to absorb/sink energy. You can set the modes the port needs to absorb on the Advanced tab of the waveguide port dialog.
  6. This is a theoretical question, beyond the scope of this forum.
  7. The out file contains all sorts of data. It might be easier to use the far field output files (*.ffe) and near field output files (*.efe/*.hfe) as these will contain only the relevant data. Then just read / combine the appropriate lines of data from these files. But if you want to write an automation script for POSTFEKO, it would be better to first start with a smaller data set: Create a smaller version of the model with say 5 configurations and 4 near field and 4 far field points per configuration and get that working first.
  8. Please attach the efe/hfe files, thanks. If you can also the model that was used to generate the efe/hfe files.
  9. If you request S-parameters, Feko will automatically load each port with 50 Ohm (or a different value if the user specifies it). You have added 50 Ohm loads manually - effectively you are calculating S-parameters in a 100 Ohm system.
  10. If you excite the TE20 mode, lower order modes will also be propagating. But it depends greatly on the waveguide dimensions and frequency. The out file will contain details of all the modes.
  11. Please share some details about what you are trying to solve.
  12. That looks like an out-of-core solution which uses the hard drive as "memory". This can be very slow. You will need to use a machine with 200 GByte RAM or more, or reduce the number of mesh elements of the model, or consider a different solver.
  13. You can set the environment variable temporarily as shown below. The rhs file is only supported for a sequential solution (parallel solving disabled).
  14. In CADFEKO you must setup the 4 patches to have e.g. 50 Ohm loads on their ports. The output request must be set to Currents. The source will then be a plane wave. Or if you have an antenna, or maybe a far field pattern source of an antenna. In POSTFEKO you plot the phase of the voltage or current by requesting the Power results. Attached some images of the setup and also the 2019.3 version Cadfeko file. The phase error graph was done with a Lua script in POSTFEKO. The script is not shareable on this forum and is part of an advanced Feko training class. The formula that was used to make the graphs is given in the below image. 4Dipoles.cfx
  • Create New...