Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Radio wave antenna calculation, solution doesn't converge

Please login with a confirmed email address before reporting spam

Hi,
I was trying to study the radiation from an nano-scale antenna as well as its impedance. However, the solution doesn't converge. I tried all iterative methods, unfortunately, none of those actually worked.
I tried to use the Gauge fixing for A-field. it gave my an error saying that "a 0 element in the matrix". I knew I may be able to fix it by increasing the conductivity of air and other high resistance components. But I don't really want to do so, because the antenna has a really low conductivity too~10^4 S/m

Any help will be appreciated.


1 Reply Last Post 14 nov. 2013, 22:18 UTC−5
Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14 nov. 2013, 22:18 UTC−5
Hello "ZQ." I make a lot of antenna models, but I've never done one that is configured the way you have configured yours. Now that, all by itself, doesn't mean you are necessarily doing it all wrong-- but it doesn't help me understand what you are doing. The first thing I notice is that you are using the magnetic-field formulation (part of the AC/DC package), while it seems to me you should be using the RF package. Hmm. Your sphere is only about 1mm in diameter. But your frequency is 2.87 GHz. Well now, your sphere is MUCH smaller than a wavelength, which is probably why you chose to use the AC/DC module, instead of the RF module. On the other hand, you are calling this a "radio wave antenna calculation." And that naming doesn't make much sense to me, frankly, since to do that right, your volume around the antenna should extend to at least a decent fraction of a wavelength (e.g., lambda/2), so that you would have some chance of providing something resembling a radiation boundary condition. I'll confess that I don't know the mathematics of the "infinite element" domain you are trying to use instead, since I haven't used those. That said, I'd be surprised if it is an adequate substitute for the aforementioned modest amount of space normally required to let your antenna actually radiate, so to speak. Then, I also notice you have this very thin (but NOT two dimensional) section driven by a sliver that is a lumped-element port. I'm guessing that might cause some numerical issues, especially since you apparently haven't made the mesh finer in the port region (note: it is generally a good idea to use relatively fine meshes on your ports). In summary, if I were doing your model, I would use the RF module, use a finer mesh on the port (and possibly even replace your thin layer with a 2D layer) and I'd use a much larger computation sphere. Oh, and you might also consider taking advantage of a plane of symmetry. Meanwhile, I encourage others here to add their comments, to tell you other places in your model where you did things wrong, and perhaps even to tell you why you shouldn't listen too much to my advice, since I really don't use the AC/DC module so much anymore... Good luck!
Hello "ZQ." I make a lot of antenna models, but I've never done one that is configured the way you have configured yours. Now that, all by itself, doesn't mean you are necessarily doing it all wrong-- but it doesn't help me understand what you are doing. The first thing I notice is that you are using the magnetic-field formulation (part of the AC/DC package), while it seems to me you should be using the RF package. Hmm. Your sphere is only about 1mm in diameter. But your frequency is 2.87 GHz. Well now, your sphere is MUCH smaller than a wavelength, which is probably why you chose to use the AC/DC module, instead of the RF module. On the other hand, you are calling this a "radio wave antenna calculation." And that naming doesn't make much sense to me, frankly, since to do that right, your volume around the antenna should extend to at least a decent fraction of a wavelength (e.g., lambda/2), so that you would have some chance of providing something resembling a radiation boundary condition. I'll confess that I don't know the mathematics of the "infinite element" domain you are trying to use instead, since I haven't used those. That said, I'd be surprised if it is an adequate substitute for the aforementioned modest amount of space normally required to let your antenna actually radiate, so to speak. Then, I also notice you have this very thin (but NOT two dimensional) section driven by a sliver that is a lumped-element port. I'm guessing that might cause some numerical issues, especially since you apparently haven't made the mesh finer in the port region (note: it is generally a good idea to use relatively fine meshes on your ports). In summary, if I were doing your model, I would use the RF module, use a finer mesh on the port (and possibly even replace your thin layer with a 2D layer) and I'd use a much larger computation sphere. Oh, and you might also consider taking advantage of a plane of symmetry. Meanwhile, I encourage others here to add their comments, to tell you other places in your model where you did things wrong, and perhaps even to tell you why you shouldn't listen too much to my advice, since I really don't use the AC/DC module so much anymore... Good luck!

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.