Jump to content

JIF

Moderators
  • Content Count

    403
  • Joined

  • Last visited

  • Days Won

    20

JIF last won the day on January 6

JIF had the most liked content!

About JIF

  • Rank
    Super User

Profile Information

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

Recent Profile Visitors

2967 profile views
  1. Hello Sem, The array feature in FEKO assumes that you are describing a unit element and thus the radome will also be duplicated. To simulate the array with the radome, you will have to create all the elements and excitations in the traditional way (not the array feature) and add the radome. If the radome is well separated from the array, model decomposition can be used (sample around the array and use that as an impressed source for the radome).
  2. JIF

    Batch Script Error

    Hi bouvy, I assume that all 40 cores are on the same machine? If so, add the following to the end of the runfeko command: --parallel-authenticate localonly
  3. Hello KateKim, A good place to start is the Getting Started Guide. ;-) There are also some patch models in the Example guide. I doubt that anyone would have time to create the models for you.
  4. Hi Could you attach the model so that we can take a look?
  5. Hi Ksewsu, I assume that you are using Ubuntu 18.04 (not 18.02 - that does not exist as far as I know). I just installed it again on Ubuntu 18.04 and it worked without issues. My guess would be that you some OpenGL problem. Try updating your graphics card driver? Any reason why you are using version 2017 and not 2018? (I'm asking since using the latest version is best, but version 2017 should work and it worked for me now - confirmed.)
  6. Hello Christopherson, You are indeed correct that this is easy. There are two ways to go about finding out how to do things in the CADFEKO API. The one is to find the object in the API and see what options it provides (properties, methods) or using the example for that object if it does roughly what you need. The other way is to use script recording - record the actions and then see the code it generated and modify it to your requirements. Below is a section that I script recorded. I simply created a plane wave source with default options and then modified it, changing the magnitude, phase and also angles of incidence. app = cf.GetApplication() project = app.Project -- Created solution entity: PlaneWaveSource1 properties = cf.PlaneWave.GetDefaultProperties() properties.Label = "PlaneWaveSource1" PlaneWaveSource1 = project.SolutionConfigurations["StandardConfiguration1"].Sources:AddPlaneWave(properties) -- Modified solution entity: PlaneWaveSource1 properties = PlaneWaveSource1:GetProperties() properties.DefinitionMethod = cf.Enums.PlaneWaveDefinitionMethodEnum.Multiple properties.EndPhi = "45" properties.EndTheta = "90" properties.Magnitude = "2" properties.Phase = "20" properties.PhiIncrement = "5" properties.ThetaIncrement = "10" PlaneWaveSource1:SetProperties(properties) If you look at the "PlaneWave" object in the API reference, you will see similar info (obviously the links won't work in my extract below): I hope the info helps.
  7. Hi This is possible and is exactly what time analysis in POSTFEKO does. The analysis is based on FFT/IFFT and thus the signal is inherently periodic (you can't avoid it).
  8. Hello RF Engineer, What seems to be the problem? What have you tried? Have you opened the Planar Multilayer Substrate dialog and clicked on the "Add" below the table of layers?
  9. Hello KacyHoppe, Was there a point or question to your statement? It does not seem relevant.
  10. Hi I agree. I'll let you in on a secret. We are doing a large refactoring (code improvement) project on CADFEKO (and after that POSTFEKO and the other applications) and this has caused many features that we know are very important to be delayed. However, once the refactoring is complete, features like this will be released soon after. Just bear with us a little longer until we are on the new (faster) code base.
  11. Hello liucass, I see that you are using an older version of Feko. Please upgrade to the latest version. The error is in Intel MPI. It sounds like you don't get it for every run, but occasionally? We have reported similar issues to Intel in the past and some of the issues were fixed by them in later versions and that is why I suggest that you upgrade to the latest version of Feko.
  12. Hello briantvd The good news is that the answer is yes, this will be improved. We are working on various features to make this faster and easier. The less good news is, these features are still in development and I'm only expecting them to become available in about a year. The features are in very early development and we can't yet guarantee that they will work as we plan for them to work, but it is being looked at and there are multiple feature that should all improve this situation. Something that you could try is to remove the pole joining the array elements to see if it has an effect on the array (pattern and impedance). If you can avoid including that pole, you don't need to union the elements and thus the problem is avoided. The problem here is the union that is taking long. When you delete an element in the union, it needs to redo the union and this is a slow process in the CAD modeller when many faces are present. The problem will get worse the more elements you add. If you have many elements or need to change the model often, it could be worth it to rather create the model using a script - it really depends on what you are trying to do. But as I said, we are aware of the problem and have solutions and have been working on them for quite a while, but will still take some time before we can release them to customers.
  13. But, CMA is a MoM solution and there is not much GPU acceleration for the MoM solver yet.
  14. Hello kevinwhite, What you use really depends on the frequency and radius in question. You can model this structure with cones that will be meshed into triangles if the radii of the elements are large (sizable fraction of a wavelength). If you are working at low frequencies and the radii are small (much smaller than a wavelength), you can use segments. EDITFEKO has a card (BL card with tapered radius) to create wires where the segments vary in diameter along the wire. The same feature does not yet exist in CADFEKO though.
  15. Hello Shane, I'm not sure what you are trying to do. Are you trying to get hold of the the current at the middle of each triangle (magnitude only, or complex or also the vector direction) for each triangle for each incident plane wave? You can get the info from the OS file or the DataSet as already indicated. The both the DataSet and the OS file provides the complex vector current at each vertex. If you want the current at the middle of the triangle, simple add the three values and divide by 3. Jc = (1/3)*(J1 + J2 + J3) Now, we can calculate the magnitude of the current at the centre of the triangle as |Jc| = sqrt( |Jcx^2| + |Jcy^2| + |Jcz^2|) = sqrt( (Re(Jcx)^2 + Im(Jcx)^2) + (Re(Jcy)^2 + Im(Jcy)^2) + (Re(Jcz)^2 + Im(Jcz)^2) ) where Jcx is the x-component of the vector Jc (complex vector current density at the centre of the triangle). Note that this is not necsesarily the maximum that the current will be, but rather the upper limit of the maximum since it assumes that the three vector components are in phase and will be a maximum at the same time. If we wanted to really calculate the maximum, then we would have to calculate the phase (wt) where the maximum will occur and then use the magnitudes of the vector components when the phase is equal to wt_max. This means that the instantaneous value calculated below might never actually be equal to the maximum calculated here. We can also calculate the instantaneous value of the current at the centre of the triangle. Jci(wt) = Re(Jc*exp(jwt)) = Re(Jc)*cos(wt) - Im(Jc)*sin(wt) Note that Jci(wt) is a scalar vector (not complex). Now, the magnitude of Jci(wt) indicating the colour in POSTFEKO is calculated as before using: |Jci(wt)| = sqrt(Jcix^2 + Jciy^2 + Jciz^2) . This explains what POSTFEKO displays in the 3D view. I hope that helps. If you do use the DataSet and not the OS file, you will see that each triangle index is repeated 3 times and this represents the current at vertex 1, 2 and 3 for that triangle. To know the exact location in space of the vertex, you need to access the mesh info (also via the API). I hope that helps.
×
×
  • Create New...