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.

Skin effect simulation in tinned/covered coil

Please login with a confirmed email address before reporting spam

Dear Comsol Comunnity,

I have been using Comsol at my University a few times so my experience in this software is only scientific and quite general. Now I have a problem which is quite complex. Here is the potential problem abstract. I want to simulate the skin effect in tinned coil or coil covered by other material like kanthal, lead, nickel and constantan. This is the air coil 20mm diameter with 14 turns with 4mm separation between them covered by around 100 um to 300um different material. It should have estimated inducted around 1 or 2uH (the exact value is doesn't matter for now). I want to do frequency sweeping from single MHz to tens of GHz for chosen material and its thickness. For example copper cover by tin 100umthickness -> frequency sweeping, same 150um, kanthal 100um then 150um etc etc. My goal is to chose the configuration where the surface current density norm (A/m) is highest. Thing is that I'm not sure how to cover the copper with other material. What I have done is STP file with example coil model I wanted to examine (attachment). I have add the boundaries for all of the coil excluding the boundries that I want to define as Coil input and output in Coil geometry analysis. My question are: 1. Whether my thinking is correct? 2. If yes (1) is this the only way to make the material coverage? 3. If my thinking is correct: a) Whether all of the boundaries which covers the copper should be defined in Coil explicit additionally or whether they should be defined as second coil and mutual inductance between them ? b) How to generate the Mesh for that kind of problem ? 4. Maybe there are special material library for that kind of problems ?

Best regards, Krzysztof Szybinski



4 Replies Last Post Jan 21, 2019, 12:29 a.m. EST
Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Jan 20, 2019, 10:36 a.m. EST

A few comments:

  • Typically you would not mesh the metal domain and use impedance boundary conditions at the surfaces instead.
  • If the coating is thinner than a few skin depths this approach gets inaccurate because current is flowing in different materials. Skin depth can be calculated analytically.
  • A frequency scan over orders of magnitude is difficult because the mesh must resolve the wavelength.
  • From basic considerations it is straightforward that the current density is the higher the higher the conductivity of the surface layer is. That is why we put silver coatings on RF components.

Cheers Edgar

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
A few comments: - Typically you would not mesh the metal domain and use impedance boundary conditions at the surfaces instead. - If the coating is thinner than a few skin depths this approach gets inaccurate because current is flowing in different materials. Skin depth can be calculated analytically. - A frequency scan over orders of magnitude is difficult because the mesh must resolve the wavelength. - From basic considerations it is straightforward that the current density is the higher the higher the conductivity of the surface layer is. That is why we put silver coatings on RF components. Cheers Edgar

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Jan 20, 2019, 1:17 p.m. EST

Dear Edgar,

Thank you for your rapid and factual answer. As for the third point. I can do in manually for example 10 Mhz 50 MHz 1 GHz etc. At the beginning I don't need to have a lot frequency measurements. As for the last point - I fully agree, but probably forgot to say that my goal is completely opposite. I want to cover the high conductive conductor (copper) by the conductor which has the lower conductivity. The think is to find the proper configuration which provide the highest losses in this frequency band. I want to use skin effect to precipitate the energy of high frequency which I want to get rid of, into heat. On the other hand I can chose the circuit excitation and connect the 3D coil into my signal (damping oscillation with very short rising time).

Wheter my coil model in STP file is correct ?

Best, Krzysztof

Dear Edgar, Thank you for your rapid and factual answer. As for the third point. I can do in manually for example 10 Mhz 50 MHz 1 GHz etc. At the beginning I don't need to have a lot frequency measurements. As for the last point - I fully agree, but probably forgot to say that my goal is completely opposite. I want to cover the high conductive conductor (copper) by the conductor which has the lower conductivity. The think is to find the proper configuration which provide the highest losses in this frequency band. I want to use skin effect to precipitate the energy of high frequency which I want to get rid of, into heat. On the other hand I can chose the circuit excitation and connect the 3D coil into my signal (damping oscillation with very short rising time). Wheter my coil model in STP file is correct ? Best, Krzysztof

Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Jan 20, 2019, 4:21 p.m. EST
Updated: 5 years ago Jan 20, 2019, 4:22 p.m. EST

Krzystof,

the coil looks ok. It took a while to import it. You could generate it inside Comsol using a spiral spine curve and sweep the cross section along it. It may be a challenge to mesh the the thin surface domain. You may try to sweep a mesh from the end surface through the coil. A boundary mesh may be another option. You need to resolve the thickness of the layer with 3-4 elements or more depending on the accuracy requirements. For your objective you will indeed need to mesh the thin layer and use an impedance BC on the inner Cu conductor.

Cheers Edgar

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Krzystof, the coil looks ok. It took a while to import it. You could generate it inside Comsol using a spiral spine curve and sweep the cross section along it. It may be a challenge to mesh the the thin surface domain. You may try to sweep a mesh from the end surface through the coil. A boundary mesh may be another option. You need to resolve the thickness of the layer with 3-4 elements or more depending on the accuracy requirements. For your objective you will indeed need to mesh the thin layer and use an impedance BC on the inner Cu conductor. Cheers Edgar

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Jan 21, 2019, 12:29 a.m. EST
Updated: 5 years ago Jan 21, 2019, 4:25 a.m. EST

Thank you very much again Edgar. This information will be the strat point for me in seting up the model.

Best, Krzysztof

Thank you very much again Edgar. This information will be the strat point for me in seting up the model. Best, Krzysztof

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.