3D Capacitance Analysis of coaxial cable using hfss

Hi there,

There might be several similar threads, but - please believe me - I couldn't find what I am searching for.

I am in a project where we use HFSS for some simulations. Since we got really screwed up results for shielded cable models (compared to none shielded ones) I was trying to just simulate the capacitance value of our cable geometry (and hopefully getting similar results than the calculated ones of the cross section values (7.9mm diameter of conductor, 10.1mm conductor + insulation, 11mm conductor + insulation + shield led to something around 370 pF/m)
So my main goal is to confirm that value by using the 3D model.
So I set up a simple model of that cable with only cylinders (open ends) of 1m length (to be exact I used CreateLine and CreatePolyline) and a radiation boundary which is a cylinder of 2m length and 1m diameter.
By that point I got stuck. How to calculate the capacitance of that cable? I didn't find a way using a voltage port (which we are using in our project) nor with a lumped port.

Thanks in advance for your help and a happy new year!

If you have designed 3D model in HFSS... you can send snap of that to understand mechanism...

Hi, this is the hfss-file to what I did to now.

What else do you need?

One_wire_shielded_capacitance.hfss.txt

Hi again,

I'm still not any further. Maybe I am trying it in a completely wrong or stupid way.
I (just) want to get the capacitance of a 1 meter coaxial cable using hfss. both ends left open. How to feed the cable? Would like to use the voltage port - but that way I wont even get the Y,Z, or S-parameters to calculate the capacity ...

I really do need help.

I get a capacitance of ~226.5 pF/meter at 10 MHz. It is of note that your model used air as the insulator, so if you insert the correct dielectric the capacitance will go up as expected. I have attached the modified project to solve this accurately.

Basic Premise:

Capacitance is equal to : C = -1/(2*pi*Freq*Im(Z(1,1))) where C is capacitance in Farads, Freq is frequency in Hz, Z(1,1) is complex self impedance of structure. It is also of note that Z is used rather than Y as Z implies that all other ports are open (which is desired) while Y implies all other ports are short (which will induce a very strong parasitic inductance to this simulation)

Have Fun

Thanks so much!

Y implies all the other ports a short. Sounds familiar (but from a far away voice in my head). :D

Will try it with Z now ... what resistance value to take for the port? Did you try it with 50 ohms?

Actually due to save mesh points we only use vacuum in our big model as well. That is why I only used vacuum.

There is no possibility to get the Z-parameters by using a voltage excitation, is it?

EDIT: Mmh .. why is that value so different for different frequencies? (even negative?)

Why would you use a voltage excitation? The port automatically determines the complex imepdance of the structure in an automated fashion... a voltage source can be used, but it would require quite a bit of fancy footwork in the field calculator. In my simulation I have no negative capacitances and the capacitance is pretty much flat except near DC?

I used a waveport, but that would not matter as the impedance matrix is completely independent of the source reference.

Have Fun

mmh strange... could you upload your file?

EDIT: sorry - was too blind to see it! My bad!

I used a lumped port with 50 ohm and the formular you gave me. Got strange results.

I cant use the wave port, cause it is within my boundary. And since I want to have the effects on both ends of the cable I dont want to have the boundary only as long as my cable!

As stated in my response, the modified project was already attached

You actually really did make some changes to that model. Since I really want to learn to handle that software (and radio wave simulations as well), could you please say one or two words on why you did those?
Port:
Why using a wave port and no lumped port or voltage exitation? I remarked that you didnt provide a integration line for that wave port at all. How is that working? Why dont you need to tell the software where your connections of the port are?
Boundary:
You are not using any boundary at all. Why dont you need one? Would it work with lets say a radiation boundary too?
expression cache and converge:
What is that? Never saw that? What is it doing? When do I need those?

I bet I could find some of the answers to my questions myself but nevertheless I would really appreciate it to get those from you!

Anyways: You helped me a lot! Thank you very much!