How to obtain properly tracked CMA modes in POSTFEKO?
Altair EmployeeAnswers
-
FEKO has greatly improved the mode tracking capabilities in the last few years and it is recommended that users use the latest version of FEKO. The tips below are still recommended for improved mode tracking.
Characteristic mode analysis (CMA) is the numerical calculation of a weighted set of orthogonal current modes that are supported on a conducting surface. This how-to explains how the mode tracking is done in FEKO and why sometimes incomplete traces for modes are obtained in POSTFEKO.
Introduction
Characteristic modes are obtained by solving a particular weighted eigenvalue equation that is derived from the method of moments (MoM) impedance matrix. FEKO has a built-in solver that calculates these modes, with no need for post-processing by the user. The eigenvalues, modal significance, characteristic angles, currents, near fields and far fields can be visualised in POSTFEKO.
However, inadequate setup of the model could result in inaccurate CMA results.
Consider a dipole constructed with a cylinder (meshed into triangles).
In CADFEKO, a characteristic mode configuration is requested and the standard configuration is deleted. The default number of modes to be calculated is kept at 6 as shown in the image.
The frequency range is set from 2.5 to 3.5 GHz. In addition, the output of currents is requested so that we can view them during the post-processing phase.
After FEKO is run, a Cartesian graph of the modal significance of the 6 modes are plotted in POSTFEKO. The graph is shown below.
We see that only mode indices 1 and 2 are complete for the frequency range. POSTFEKO also shows warnings for traces 3 to 6. The reason for the incomplete traces for the other modes is as follows:
When a frequency range is requested, mode numbers (rather indexes) are assigned according to their ranking at the lowest frequency. This initial ranking or sorting is based on the modal significance of the mode, that is, modes with a smaller eigenvalue will have a lower ranking. FEKO tracks the requested number of modes over the complete frequency range.
If the user requested, for example, 6 modes, more modes are actually calculated to track every mode index over the whole frequency range. As the frequency changes, mode rankings will change. It could be that a mode that had a very low eigenvalue at the lowest frequency, and that was assigned a very low ranking, changes to a high eigenvalue at a higher frequency. FEKO will order the modes correctly at each frequency, but this ordering will be based on the mode rankings at the lowest frequency.
Now since many more modes are to be calculated, these higher order modes often require a much finer mesh than the well known 1/10th of a wavelength for a MoM solution. If the mesh size is not sufficiently small for a particular mode, the mode cannot be represented accurately by the mesh and the mode tracking will fail. The mode tracking could fail at any frequency over the range. Consequently an incomplete curve could be obtained for a mode in POSTFEKO.
If we look in the result palette in POSTFEKO in the Mode index dropdown, we see several modes listed, although we only requested 6 modes:
This is also because a mode could not be tracked over the entire frequency range. For example a mode could have been assigned an index at the start frequency, but at a higher frequency this mode could not be tracked any more and was assigned a new index. It could even happen that the 'new mode' is also untracked at a higher frequency, and again assigned a new index.
The remedy is to use a finer mesh. The mesh size is adjusted using a 3 times smaller triangle edge length and the model is rerun.We see in the graph below all the modes are now tracked over the complete frequency range:
In addition, if we expand the result palette again we now see only 6 modes listed:
Final remarks
The requested frequency range for this example was 21 frequencies from 2.5 to 3.5 GHz. In some cases, it could also be required to use finer sampling (more frequency samples) to obtain proper mode tracking. Continuous frequency sampling is not yet supported for CMA.
0 -
I have a same problem. After I configuration mesh with fine mesh but after the simulate has been finished. The CMA cannot track along the frequency and discontinuous. Moreover, the notice id is 37561 'CMA may require a finer mesh compared to standard full-wave solution techniques, to accurately capture HOM behavior'. How can I solved this problem ?
In addition, frequency sample is 31 points from 6-9GHz and CMA index is more than 6 index.Best regards,
Thank you.
<?xml version="1.0" encoding="UTF-8"?>
.png.81ca78d3187b1e006f4f5d20bd52ff75.png)
0 -
A finer mesh could resolve the problem (as suggested in the warning). Note that you can also set a custom mesh setting that is finer than the fine auto meshing option. I would also not recommend setting the curvature refinement settings all the way to the left (or right) - having these settings on the extreme could lead to extreme meshing conditions that you may not want in your model.
I suspect the biggest problem is that you have 31 point from 6 to 9 GHz. This is a point every 100 MHz. In the comment above (Final remark) 21 frequencies were used over a 1 GHz bandwidth and it was indicated that this could be increased further to ensure good tracking. I would thus recommend that you increase the number of frequency points first and then consider a finer mesh (custom mesh setting based on wavelength). Without the model, I can't make any further suggestions.
0 -
Thank you for your reply.
In this case, can you suggest the suitable of number of frequency points ?
Best regards,
0 -
It is problem dependent, but I would use at least 61 points over the 3 GHz range and possibly even more. It really depends on the problem and the resources that you have available - more frequency points means longer simulation time.
0 -
I have followed your suggestions for CMA configuration. However, the CMA can not track along frequency as shown in the fig. below. Could you solved this problem ?
In addition, I have attached file below.
Sorry about my FEKO skills.
Best regards,.png.0afafda8bcc74180b77375c0b32a45ff.png)
0 -
Hi BM Kung,
I ran your problem with your original proposed 31 frequency points over the 3GHz range and there are no tracking errors, even with 100MHz steps.
I think what you are having a problem with is mode 4 and mode 6 not being reported over the whole frequency range?
What is happening here is that FEKO solves the CMA problem independently at each frequency point and returns the most significant modes up to the number of requested modes (in this case 6). Tracking is then applied as a post-processing step over frequency samples. If a mode that is initially ranked within the top 6 loses its ranking at a higher frequency FEKO will stop tracking that mode and return the new more significant mode instead.0 -
HI Wilco Strydom,
Yes my problem is how to configuration to solve the CMA tracking for mode 1- 6 ?
I try to increase frequency points up to 301 points (2 GHz - 5 GHz) and use local mesh setup of Lambda/20.
But the CMA still tracking incorrectly.0 -
With a request of 6 modes FEKO will give the most significant 6 modes at every frequency. It is not a tracking error when modes fall out of significance and are replaced by others.
If you would like the track the 6 most significant modes at the first frequency across the whole frequency range you will need to request more modes (10 should be sufficient) to account for the fact that some of those 6 modes are likely to drop in significance.0 -
Hi BM Kung
You can increase the number of characteristic modes to be calculated where you set up the problem in CADFEKO:
I reran your model with 12 modes requested and then the first 6 modes were completely tracked across the frequency range from 2 to 5 GHz. (You could perhaps get the results you require by requesting fewer than that, i.e. selecting between 6 and 12 modes.)
0 -
Can you suggest mesh size?
0 -
Can anyone please share meshing for the multilayered structures
like top layer is antenna, 2nd layer is substrate, 3rd layer is air and 4th is ground plane.
0 -
increasing mesh or no of CMS desnt solve the tracking problem of CMs. Pls clarify
0 -
Usually in patch antennas. mode 1 is always dominant bt stangely i am not getting that and its shown untracked. Pls explain.
0
