FEKO Wire port: "Segmentation rules at a connection point have been violated"; "Undefined wire port segment". How to identify the location and fix?

P.S. the inner conductor is a tube because I do not know a way to assign a non-perfect metal to a cylinder.

I am exhausted. This crippling wire port does not work no matter how I try. See the images, there I made port-adapters to the cable with two flat rings of PEC, and the central silver tube just extended. These object-origin vertices are connected by the port wire, so it SHOULD work even with "bad" meshing.

The following message from the master process (MYID= 0):
The following message from the server process (MYID= 2):
ERROR 16: Undefined wire port segment
See also message in the output file Coax1a.out
The following message from the server process (MYID= 1):
ERROR 16: Undefined wire port segment
See also message in the output file Coax1a.out
The following message from the server process (MYID= 3):
ERROR 16: Undefined wire port segment
See also message in the output file Coax1a.out
ERROR 16: Undefined wire port segment
See also message in the output file Coax1a.out

BTW, the message is wrong. The file was saved as Coax1b.cfx, but the error still was throwing Coax1a.out

I am exhausted. This crippling wire port does not work no matter how I try. See the images, there I made port-adapters to the cable with two flat rings of PEC, and the central silver tube just extended. These object-origin vertices are connected by the port wire, so it SHOULD work even with "bad" meshing.

The following message from the master process (MYID= 0):
The following message from the server process (MYID= 2):
ERROR 16: Undefined wire port segment
See also message in the output file Coax1a.out
The following message from the server process (MYID= 1):
ERROR 16: Undefined wire port segment
See also message in the output file Coax1a.out
The following message from the server process (MYID= 3):
ERROR 16: Undefined wire port segment
See also message in the output file Coax1a.out
ERROR 16: Undefined wire port segment
See also message in the output file Coax1a.out

 

BTW, the message is wrong. The file was saved as Coax1b.cfx, but the error still was throwing Coax1a.out

WAAARGH (T_T)

In the Wire Port replaced the "Segment" with "Vertex" and got:

The following message from the server process (MYID= 2):
The following message from the server process (MYID= 3):
ERROR 1161: No triangles/segments found with the correct label to apply the skin effect approximation
The following message from the server process (MYID= 1):
ERROR 1161: No triangles/segments found with the correct label to apply the skin effect approximation
ERROR 1161: No triangles/segments found with the correct label to apply the skin effect approximation
See also message in the output file Coax1b.out
See also message in the output file Coax1b.out
See also message in the output file Coax1b.out

And the content of the "out" file:

Altair Feko - Solver (par) Version 2024-4313 from 2024-04-19           Altair Feko - PREFEKO Version 2023-2564 from 2023-05-25 [FEK: default 195; used 195] BINARY           CADFEKO version 2024-4313 from 2024-04-19 . . .           Scaling of the geometry with factor  1.00000E-03                      Linear segments are used            Length of the segments in m:         1.84000E-03 No segment with the label Union1.Wire2.WirePort1 has been found  ERROR      16: Undefined wire port segment  Aborted: 2024-06-12 13:58:26 

And no, there are no such wire as Union1.Wire2.WirePort1 in the model left after the first attempt(the post above) was redesigned. It was deleted with all the references, and the engine allowed the deletion.

WAAARGH (T_T)

 

In the Wire Port replaced the "Segment" with "Vertex" and got:

The following message from the server process (MYID= 2):
The following message from the server process (MYID= 3):
ERROR 1161: No triangles/segments found with the correct label to apply the skin effect approximation
The following message from the server process (MYID= 1):
ERROR 1161: No triangles/segments found with the correct label to apply the skin effect approximation
ERROR 1161: No triangles/segments found with the correct label to apply the skin effect approximation
See also message in the output file Coax1b.out
See also message in the output file Coax1b.out
See also message in the output file Coax1b.out

And the content of the "out" file:

Altair Feko - Solver (par) Version 2024-4313 from 2024-04-19           Altair Feko - PREFEKO Version 2023-2564 from 2023-05-25 [FEK: default 195; used 195] BINARY           CADFEKO version 2024-4313 from 2024-04-19 . . .           Scaling of the geometry with factor  1.00000E-03                      Linear segments are used            Length of the segments in m:         1.84000E-03 No segment with the label Union1.Wire2.WirePort1 has been found  ERROR      16: Undefined wire port segment  Aborted: 2024-06-12 13:58:26 

And no, there are no such wire as Union1.Wire2.WirePort1 in the model left after the first attempt(the post above) was redesigned. It was deleted with all the references, and the engine allowed the deletion.

OK. I redefined the VoltageSource wich was referencing WirePort1, and made the same reference... but the "Union1.Wire2.WirePort1" error disappeared - so the UI incorrectly reports the reference to the user as a port name, ignoring tha fact that it is a wire name. And the wire name dependency is not checked correctly.

Now I have the same ERROR, but with the different "OUT" log:

Spice Engine:  HyperSpice Version 2024.0-2f4fd98a           Scaling of the geometry with factor  1.00000E-03           Surface of all triangles in m*m:     2.16455E-05                     Linear segments are used           Length of the segments in m:         3.68000E-03                                  DATA FOR MEMORY USAGE ...<skipped>  read from input file: ...<skipped>        ==============       CONFIGURATION: StandardConfiguration1       ==============                                     DATA FOR DIELECTRIC MEDIA ...<skipped>                      EXCITATION BY VOLTAGE SOURCE AT A SEGMENT  Name:                          VoltageSource1  Source index:                  1 ...<skipped>  Port at segment with label:    Union1.Wire2.WirePort1  Absolute number of segment:    2  Radius of segment in m:        5.00000E-07  Location of the port in m:     x =  0.00000E+00, y = -5.00000E-04, z =  1.73500E-03  Positive direction:            x =  0.00000E+00, y =  0.00000E+00, z =  1.00000E+00                           DATA FOR THE GREEN'S FUNCTION                                    free space                       DATA FOR LABELS  Triangle number:   77  Thickness:          8.15000E-04  Lateral dimension:  3.94897E-04  WARNING  2757: A triangle is too thick as compared to the lateral dimensions for application of the skin effect approximation Received message from server process  3:   ERROR    1161: No triangles/segments found with the correct label to apply the skin effect approximation  Aborted: 2024-06-12 14:09:01 Received message from server process  2:   ERROR    1161: No triangles/segments found with the correct label to apply the skin effect approximation  Aborted: 2024-06-12 14:09:01 Received message from server process  1:   ERROR    1161: No triangles/segments found with the correct label to apply the skin effect approximation  Aborted: 2024-06-12 14:09:01 

What does that mean!?

Please help, it drives me crazy and I am loosing valuable time!

Can you please see what is wrong in the attached project? I am completely lost , and cannot find any reason to such errors.

Hi Andrew,

I will have a look at the model today. Maybe there is an extremely small piece of wire sticking out (just a guess)?

Best regards,
Torben

Hi Andrew,

I don't know why, but when I force CADFEKO to remesh, it works. I did this by changing the mesh settings from coarse to standard and then back again. Feko has a new mesh engine implemented and I assume that there are still teething troubles...

Try attached Coax1b_alt.cfx.

Best regards,
Torben

Hi Andrew,

I don't know why, but when I force CADFEKO to remesh, it works. I did this by changing the mesh settings from coarse to standard and then back again. Feko has a new mesh engine implemented and I assume that there are still teething troubles...

Try attached Coax1b_alt.cfx.

Best regards,
Torben

Yesterday I avoided this problem by cutting off the central silver core and putting there a PEC "adapter" ring, the same as at the shield. And it worked.

I thought, the problem is in applied material at the port connection. But I am surprised that it worked in your case as is (with remeshing)

Hi Andrew,

I don't know why, but when I force CADFEKO to remesh, it works. I did this by changing the mesh settings from coarse to standard and then back again. Feko has a new mesh engine implemented and I assume that there are still teething troubles...

Try attached Coax1b_alt.cfx.

Best regards,
Torben

But I am still haunted by different values of Impedance and S11 (consiering S11should be zero when Z=50) in Optimizer and POSTFEKO; in S11 request, and Port power source request. 

As you see I wanted to search for appropriate epsilon.

When fitting Zin to 50 ohm(as the port) it goes to Eps=2.7 or above.

When fitting to S11->0, ot converges at eps=2.3

But when searching to VSWR<1.15, it converges to epsilon=1.83

And when plotting optimizer S11 as a cost function, it is also different from the requested S11 vs Freq. in POSTFEKO.

I am exhausted.

The Coax network and wave impedances should be equal, it is not an RWG.

Something wrong with the port. But it is the only port with controllable impedance which may be used in the optimizer as a cost.source and as power source for radiation simultaneously.

※BTW, is the goal in fact the ERROR function ? I.e. for S11->-30dB ERROR=S11-(-30)=optimization.goals.sparametergoal ?