nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Siemens Community Catalyst Program

The Siemens Community Catalyst program was co-created with our community to acknowledge technology leaders who consistently contribute to the Siemens Community. Nominations are accepted on a rolling basis.

Nominate Now

Verify Shunt Capacitor Sweep Setup in FEKO

DanielPerk

I'm looking to plot S-Parameters for the purpose of extracting capacitance values throughout a frequency sweep. I am able to run the simulation, plot S-parameters and export the SnP file, but here are the issues:

Either the capacitance is about twice what it should be, or
this shunt capacitor electrically looks like a series inductor

My setup is very simple:

One outer cylindrical shell with ID = 6.00 inches, wall thickness = 0.125 inches and width = 8.00 inches
One solid cylindrical inner conductor with OD = 5.00 inches, width = 7.50 inches.
primitive cylinder surrounding all geometry with about 8 inches of padding

I assign "air" (dielectric that I created) to the larger encompassing primitive cylinder, and the other two bodies are perfect electrical conductors. I added two lines, each starting on the inner conductor and ending one the outer conductor—one on each side of the capacitor. I union all solid bodies and lines, then assign wire ports to each line and put the "location" in the middle of the lines.

Next, i request S-params and specify both ports as active, provide a frequency range to sweep and run.

I'm not sure what I could be doing wrong here, so if someone could please take a look at my setup, I would appreciate it. The calculated capacitance for this geometry should come out around 60 pF, but I am getting 180 pF.

I've uploaded the .cfx file.

Thank you,

Daniel Perk

8in air cap FEKO sim.cfx

Find more posts tagged with

Feko

Accepted answers

DanielPerk

Jaehoon,

Thank you for your response and time. The model runs fine and results are as expected!

I'm curious, is the only thing you changed about my model was change the region (the one that is available only when "union" is selected) from pec to fep?
Its tricky, because I brought in three solids and properly designated them as pec or dielectric, but when i unioned the cylinder with the three bodies, there is another region that needs to be designated as dielectric.

Thank you again for your help Jaehoon.

Are there any other ports that would work for this simulation aside from waveguide port?

Jaehoon

Hello Daniel,

I think the symmetry issue is mostly due to geometry point mismatches (This happens sometimes when importing CAD). For example, I selected two points, whose z values are different. Feko thinks the associated edges are not equivalent.

To ensure the symmetry, I used the Feko "Split", "Delete" and "Copy (mirror)" two times to extend a quarter of the geometry for the whole geometry construction, because there are 2 different symmetries. (Please see the attached model.)

4-5-in-air-cap-symmetry-issues-feko_mod.cfx

Please make sure to assign proper region materials by using the cut-planes.

Best regards,

Jaehoon

All comments

Jaehoon

Hello @DanielPerk ,

I'd like to raise some questions to understand your simulation's target geometry.

Do you want to extract the shunt capacitance value of the general (normal) coaxial cable (whose dimension you shared already, normally two cylinders) or the exact geometry (as below, there are three cylinders) in your model?

If your answer is a general coaxial cable, then I will suggest other way or modify your model (like feeding structure).

If you answer is the exact geometry in the model, please let me know how much I can modify your model for more accurate results? (We don't normally use a wire for Coaxial cable feeding.)

Best regards,

Jaehoon

DanielPerk

Hello again Jaehoon,

I need the exact geometry, as I will have other models that have hump in the center where the two conductors come into close contact and sandwich a dielectric. The third cylinder is air, and I changed it's relative permittivity to 2 just to see if it was being taken into account.

DanielPerk

8 in Cap with 2 in hump_FEP_FEKO.cfx

This is the geometry I referred to with the internal "hump". The dielectric has rel permittivity of 2.1.

I was working with the air capacitor first because it has simple geometry. Hand calcs show the air cap should have a capacitance of around 60 pF. My goal was to run the air cap, export S-parmeters and extract capacitance that equaled somewhere around 60 pF. Does this make sense?
Then, if I got that to work, I was going to move on to the more complicated geometry I have attached here.

Jaehoon

Hello Daniel,

Please have the attached model modified by Feko 2026.

8in-air-cap-feko-sim_mod.zip

I did the followings..

drew the outer conductor again
gave the dielectric medium (aira) to the region between the inner and outer conductors.
defined two waveguide ports which are more natural for the coaxial cable feeding.
checked the double precision
activate two magnetic symmetries.

Best regards,

Jaehoon

DanielPerk

Jaehoon,

Thank you for your assistance. The Smith chart trace looks like it should and extracting capacitance from the Y parameters results in a capacitance of 62 pF at 40 MHz—hand calcs give 59 pF.

I believe this same setup will work for the second model, so I will mimic your approach on the other geometry and go from there.

Thank you again for your help!

Daniel

DanielPerk

Jaehoon,

If you get an opportunity to review the second .cfx I sent, please tell me how you would go about creating the outer conductor for that geometry. I will upload my attempt in the next message…

Jaehoon

Hello Daniel,

Let me explain what I did for the outer conductor in the modified model.

I used *Separate" and "Disassemble" features a few times and deleted your original outer conductor geometry after measuring the sizes of the original (radius and length). (You can check them in the model.)

And I constructed one same-sized cylinder and union(ed) all the geometries..Then the region for the dielectric and PEC are automatically separated.

Best regards,

Jaehoon

DanielPerk

8 in Cap with 2 in hump_FEP_FEKO.cfx

Jaehoon,

This is my setup with more complicated geometry. I did run it, but the smith chart and freq/magnitude plot both looked unusual. In this model, I removed the outer shell, similar to what you did in the first model. Unfortunately, there is a bit more geometry associated with the outer conductor. Additionally, the inner cylinder is wrapped with a dielectric.

I followed your setup, but where the two parts meet and share a wall/surface, im not sure if I should make those walls part of the PEC or of the dielectric. Please assist, and thank for you help.

Jaehoon

Hello Daniel,

Your waveguide source setting looks OK.

I am not sure what is the role of the two metallic rings between the inner and outer cylinders.

I mean all the power from the source should be reflected back due to the rings.

Best regards,

Jaehoon

DanielPerk

Jaehoon,

The full geometry is a coaxial capacitor. Below is an image of the full geometry. the "rings" you asked about were part of the outer conductor.

The "rings" are there to increase capacitance, and the blue body is a dielectric that can withstand the Efield that is present with such close proximity inner/outer conductors.

In the .cfx I sent, I deleted all the outer conductor except for this ring, and replace the deleted part with the FEKO created cylinder (to follow your setup).

At 1–40 MHz, that hump will cause some localized impedance mismatch and partial reflection, but nowhere near “all power reflected” unless the discontinuity is extremely severe or the structure hits a resonance mode. More likely, it creates localized field enhancement, parasitic resonance, and some degradation in return loss/S11 instead of acting like a total RF wall.

DanielPerk

If using a waveguide port will result in 100% of the power being reflected, then is there a better port to model this geometry. In my lab, this exact geometry is built and we test it in filters. Energy definitely goes through the filter. As you are aware, filters require capacitors with specific capacitance, and this geometry provides values we need. Being able to accurately model this geometry and output s-parameters is vital to our work.

Jaehoon

Hello Daniel,

Thank you for the detailed explanation which is helpful for me to understand your simulation geometry.

Based on that, I modified your 2nd model and attached. (The region between the ring and the inner conductor should be a dielectric…I am not sure FEP is a right material.)

8-in-cap-with-2-in-hump-fep-feko_mod.zip

Best regards,

Jaehoon

PS) You don't have to change the port..

DanielPerk

Jaehoon,

Thank you for your response and time. The model runs fine and results are as expected!

Thank you again for your help Jaehoon.

Are there any other ports that would work for this simulation aside from waveguide port?

Jaehoon

Here are some answers..

Q: is the only thing you changed about my model was change the region (the one that is available only when "union" is selected) from pec to fep? Yes, that's right. It is safe to check the material properties after union.

Q: Are there any other ports that would work for this simulation aside from waveguide port? Possible alternative is the wire or edge port, but the waveguide port is more natural (=free from a transition) than any other.

Thank you,

Jaehoon

DanielPerk

Jaehoon,

I've started running various capacitors with this same setup that you've helped me with and have ran into something I cant quite figure out…

When I apply symmetry on one of my models, despite the geometry being symmetrical, as all the other capacitors have been, I get the message that 10+ edges are not symmetrical.

Upon inspection, the edges (and faces) when imported, get bisected or sometimes split into quarters. This may have something to do with the lack of symmetry, but I'm not sure. Did you have to edit the geometry or change anything to get symmetry to work with the geometry I've thus far provided?

Jaehoon

Hello Daniel,

Strange..The last model (8-in-cap-with-2-in-hump-fep-feko_mod.zip) I shared with you has two magnetic symmetries with no issue as shown below. (I didn't do anything from your model for the symmetry.)

Please check the model again.

You might run the simulation with the symmety..requiring more time, though.

Best regards,

Jaehoon

DanielPerk

Jaehoon,

Thanks for the response. When adding symmetry in other simulation software, I would only need to model/mesh 1/4 or 1/4 of the geometry and apply symmetry (where applicable) to run faster.

So is that the only purpose of adding symmetry here in FEKO, to decrease run time?

Jaehoon

So is that the only purpose of adding symmetry here in FEKO, to decrease run time?

→ Yes, that's right.

Please share your model..Let me check…

Best regards,

Jaehoon

DanielPerk

4_5 in Air Cap_Symmetry Issues_FEKO.cfx

I've uploaded a .cfx for a simple 4.5 in air cap that doesn't like symmetry. The model was constructed the same way as those other 8 in caps we worked on, so it was modeled about the center-planes with the origin located at the cog of the symmetrical parts.

Please let me know why symmetry cannot be added here.

Jaehoon

Hello Daniel,

4-5-in-air-cap-symmetry-issues-feko_mod.cfx

Please make sure to assign proper region materials by using the cut-planes.

Best regards,

Jaehoon

DanielPerk

Got it. Thank you Jaehoon!