Discussion Forum

Posted: 1 decade ago 18 oct. 2010, 02:53 UTC−4

Hi I do not fully catch your model, you have only one input port, no ground, and for me a dipole has (as its name says) two poles, but I could have missed something too -- Good luck Ivar

Posted: 1 decade ago 18 oct. 2010, 05:49 UTC−4

Hi, I only worked out how to implement core ideas of old tutorial, in V4. A lumped port is equivalent to a circuit or cable, that I used. I do not think that a separate ground connection is needed to drive a dipole in COMSOL. Actually I want to know how to plot farfield pattern, without requesting over the sphere, the doughnut is an example of dipole. you may give any other antenna example, that shows pattern in conventional / physical way I need both 3D and 2D plots. Thanks Irfan

Posted: 1 decade ago 4 juil. 2011, 08:46 UTC−4

Hi, I have the same problem. How to plot the radiation pattern? Any hint??? Cheers, Khalid

Posted: 1 decade ago 4 juil. 2011, 09:01 UTC−4

Hi, you must add a Far Field Calculations Node under the electromagnetic waves (emw) node. This will make the far field variables available for postprocessing. I advise only to use 2D plots, e.g. for azimuth and elevation patterns. 3D plots take ages to calculate, on my machine they take more time than solving the model. Cheers Edgar

Posted: 1 decade ago 4 juil. 2011, 09:14 UTC−4

Hi Edgar, I downloaded your model ( dipole 3D) . You added two slices. Are they the 2D Plots that you mean? If so, how can i calculate the gain and the 3-db beamwidth of the same model? It will help much . Cheers, Khalid

Posted: 1 decade ago 4 juil. 2011, 09:36 UTC−4

No, that is an earlier model. I am attaching the model of a 4 element yagi. Look at the polar plots. I needed to clear the solution, otherwise the model would be too big to upload. Cheers Edgar

Attachments:

• 4el_UHF_PML_cleared.mph

Posted: 1 decade ago 4 juil. 2011, 09:48 UTC−4

Hi Edgar, I can't open the file. This message is shown: " COMSOL Error File Contains no model" I am using version 4.0.0.925 Cheers Khalid

Posted: 1 decade ago 4 juil. 2011, 09:58 UTC−4

It is a version 4.1 model.

Posted: 1 decade ago 6 juil. 2011, 20:42 UTC−4

Hi Edgar, This model is very useful. And the polar plot is very fast. Could you explain why you used the expressions in the angle data section. And should i use the same expressions for any model when ploting the far field pattern? Cheers, Khalid

Posted: 1 decade ago 7 juil. 2011, 05:08 UTC−4

Khalid, the expression gives you the angle in the polar plot and allows to rotate the pattern. Arc length works as well but the results are rotated by 90 degrees. I wanted the main lobe point to 0 degrees. Just a matter of taste (or definition). Cheers Edgar

Posted: 1 decade ago 22 août 2011, 13:26 UTC−4

[QUOTE] No, that is an earlier model. I am attaching the model of a 4 element yagi. Look at the polar plots. I needed to clear the solution, otherwise the model would be too big to upload. Cheers Edgar [/QUOTE] Hi Edgar, I've just started working on COMSOL , and i needed a little help to understand your model . It would really help me a lot if the doubts are clarified. 1. In the type field for scattering boundary condition you have used plane wave but for the PML you have used Spherical.Shouldn't the type for the scattering field condition be spherical like the PML? 2.Is it necessary to have scattering boundary condition when there's a PML condition, as the primary purpose of using scattering boundary condition is for absorption of incoming waves which the PML region does effectively. 3. What's the basis for using Zref=28ohm , and in the impedance boundary condition the value 3.774e7 is based on which material. Lastly will a 64 bit windows OS with 8 gb ram not be sufficient to run the 3D radiation pattern? How long might it take approximately ? Thank a Lot in advance , your model and comments have helped me tremendously! Looking forward to hear for your reply. Regards Glenston

Posted: 1 decade ago 23 août 2011, 06:37 UTC−4

Hi Glenston, this model was a first attempt to try far field calculations in RF. I haven't been doing much more about it recently and there may be some inconstistencies. But I will try to answer as far as possible: 1. and 2. I think you are right, the PML should be sufficient. 3. The 28 Ohms is a measured value from a realized yagi antenna of this kind and also from EZNEC, that is a momentum based antenna simulator. 4. The impedance boundary value may be the conductivity of aluminium or copper, not sure about that. I ran the model on a 64 bit windows 7 machine with 12 GB RAM. For one frequency it may take some 10-15 minutes to solve and 20-30 minutes to calculate the far field in 3D. The 2D patterns are much quicker. I think 8 GB should be sufficient too. For a first test, disable the parametric sweep and solve it for one frequency. Regards Edgar

Posted: 1 decade ago 23 août 2011, 06:56 UTC−4

Hi Edgar, Thanks a lot for your patience and reply . It will help me a lot. Another totally different question. Ive built a parametric curve in 2D geometry with specifications as follows. parameter name :s range from 0-20. x exp: (125s+700)*cos(s) y exp: (125s+700)*sin(s) This gives me a 2 arm spiral around (0,0) with negligible width. I want to have a finite width to this spiral in 2D. Do you know how i can do that? Thanks a lot again for your time and effort! Regards Glenston

Posted: 1 decade ago 23 août 2011, 08:35 UTC−4

Hi Glenston, I am not completely sure what you mean. A curve doesn't have a width. If you want to create a spiral geometry (a spiral antenna?) you might consider two spirals that are rotated a bit with respect to each other and use the curves as boundaries for the spiral domain. Regards Edgar

Posted: 1 decade ago 23 août 2011, 09:33 UTC−4

[QUOTE] Hi Glenston, I am not completely sure what you mean. A curve doesn't have a width. If you want to create a spiral geometry (a spiral antenna?) you might consider two spirals that are rotated a bit with respect to each other and use the curves as boundaries for the spiral domain. Regards Edgar [/QUOTE] Dear Edgar, There's an option in CST for the width ,when you execute an equation based geometry. So I was wondering if there's a similar option in COMSOL. I'm attaching the geometry that I have created , its for a 2 arm spiral archimedian antenna. You're suggesting on creating 4 different parametric curves? Thanks a lot for your time and effort!!!! Regards Glenston

Attachments:

• case1.mph

Posted: 1 decade ago 24 août 2011, 10:00 UTC−4

Hi Edgar, I understood how to create a spiral antenna. Thanks a lot for your help. Regards Glenston

Posted: 1 decade ago 6 sept. 2011, 14:43 UTC−4

[QUOTE] Hi Edgar, This model is very useful. And the polar plot is very fast. Could you explain why you used the expressions in the angle data section. And should i use the same expressions for any model when ploting the far field pattern? Cheers, Khalid [/QUOTE] Hi Khalid How you opened the Edgar's Model I could not open it, I am using 4.2 Thanks Irfan

Posted: 1 decade ago 7 sept. 2011, 04:38 UTC−4

Irfan, it is a version 4.1 model. I recently updated to 4.2 and can open it without problems in 4.2 Cheers Edgar