Curved IFA efficiency calculation doesn't seem to match S11 results

[TGarAuk]

Hi, I am trying to simulate a curved IFA on a circular PCB with a resonant frequency of 2.45GHz. I used the FreeCAD-openEMS-exporter tool to generate the model and used their IFA example as reference.

Curved IFA design:

I am able to simulate the antenna successfully and the S11 performance is showing -36dB at the resonant frequency (2.466GHz). I also wanted to see the antenna efficiency to I copied the efficiency calculation code from the packaged inverted_f.m example (https://github.com/thliebig/openEMS/blob/master/matlab/examples/antennas/inverted_f.m).

However, while the S11 and impedance graphs show a close match, the efficiency is only ~45%. Meanwhile, the IFA example has efficiency of 81% and S11 magnitude of -31dB.

Curved IFA results:

IFA example results:

What is could be causing the low efficiency result on my curved IFA simulation?

For reference I am using:

• Windows 11
• Octave 10.1.0
• openEMS-0.0.36
• FreeCAD 1.0.0

Files:
Curved_IFA_openEMS_simulation.zip

Thanks in advance for your help.

[LubomirJagos42]

Hi,
I looked at your model and efficiency is low because you have to move antenna fine mesh to have higher priority than gnd and make mesh more fine to connect it with top gnd since now it's disconnected when you open debug PEC in paraview.
This is your coarse mesh where inverted F is disconnected from top gnd.

So I just increase mesh lines approximatiley 3x to make mesh finer

Then after simulation I got results:

##finer mesh, both top and bottom gnd used
radiated power: Prad = 7.4226e-25 Watt
directivity: Dmax = 2.038 (3.092 dBi)
fficiency: nu_rad = 99.6387 %

But thing is that you put there top and bottom gnd but inverted f antenna doesn't use bottom gnd ONLY TOP, so this are result when I commented bottom gnd and added 3 lines in Z for substrate

##corrected model only top gnd
radiated power: Prad = 4.0882e-25 Watt
directivity: Dmax = 2.0391 (3.0943 dBi)
efficiency: nu_rad = 93.386 %

So I did:

• finer mesh for antenna shape 3x
• removed bottom gnd
• added 3 lines in Z for substrate

Attaching modified script.
Curved_IFA_openEMS_simulation.zip

[LubomirJagos42]

if this solved your problem please close this issue to have it properly categorized on this forum (there should be button which to close which is possible just to author of issue), thanks

[thliebig]

@LubomirJagos42, thank you for looking into this. This is a discussion not an issue. Thus I think it need not be closed? Not sure what that even means for a discussion.

[LubomirJagos42]

:D me neither, for me it's just like habit from work to close tickets to just mark it for future that question was somehow ended and there is some resolution

[LubomirJagos42]

Hi, for curiosity I tried to rework your model into cylindrical coordinates since I never really used them in openEMS and I made some support for them into my addon but never really tested them on many examples.
So I little rework your model to mud input trace for inverted F antenna to lie directly on radial gridlines and tried to export simulation script from my freecad addon what doesn't work properly (so cylindrical coord model simulation script must be corrected) so I modified it manualy to got right results.
Attaching my simulation script if somebody is interested how to do simulations in cylindrical coordinations. For gridlines I still used in this script mesh.x, mesh.y, mesh.z what is in cylindrical coords mesh.r, mesh.theta, mesh.z

FreeCAD model:

Mesh:

Here are results:

Simulation files below.
antenna_ifa_s11_openEMS_simulation.zip

[LubomirJagos42]

I made that working with NF2FF also, there are some bugs in AppCSXCAD probably:

• when NF2FF box defined in cylinder coords in AppCSXCAD there isn't shown top and bottom 'lid' (face) of cylinder around antenna

Also when --debug-PEC used then exported PEC mesh is not properply displayed in paraview since it's using cylindrical coords and not cartesian :(

Anyway, here is simulation with running NF2FF properly attached.
Curved_IFA_openEMS_simulation_with_NF2FF.zip