HFSS simulations probleem - help needed

HFSS never again

I don´t know where the problem is. I´ve a model fast to simulate (5 minutes), it converges in a few iterations and it´s very simple to understand and draw.
I´ve built it in the reality and calculated the radiation peaks.

Right now, everything sounds good, but the results vary enormous from simulation to simulation. I think I´ve not seen two times the same S-Parameter graphic. I tried and reviewed EVERYTHING (PML, boundary radiation condition, different sizes, more strict mesh...) and I have no more ideas.

I´m student, and I should present this simulations in two days. But right now I have no consistent result to show. I will never recommend HFSS to nobody. I´ve done all the tutorials to learn, and I think I know HFSS quite well after lots of hours (I still have a lot to learn), but I cannot simulate a model as simple as mine.

HFSS never again.

HFSS never again

The best would be if you upload you problem-project and describe what you expect from it. Otherwise it is difficult to evaluate what is going there wrong.

Please upload. HFSS is very, very dependable if you truly know what you're doing with it. Without the project, sounds like it's not truly converged. How many consecutive converged passes do you require? It should be more than 1.

I really appreciate your help. I´ve uploaded my last model.

If the varialble "Wire_Turns" is set to 16, the result should be the attached in the image. This is the real S11 parameter of the prototype. As you can see, the results are absolutely differents.

With less than 4-6 loops ("Wire_Turns project variable"), the results converges fast, in only 5-10 minutes and 5-6 passes. I´ve set the convergence in 0.02 and a maximum of 20 passes. The convergence seems to be good.

I´ll really appreciate any advice. Thanks for your help.

Without running the simulation, your adaptive solution frequency needs to be the upper end of your sweep range at minimum. This is a big problem. Your mesh isn't fine enough for the upper end of frequencies, or really anything much above 1.2 GHz. When I changed the solution frequency to 10GHz, HFSS popped up a message telling me that an extremely large mesh would be generated - I'm guessing you didn't get this.

If you can get it to run, I'd also recommend changing the minimum number of converged passes to something other than 1. This tells how many consecutive passes with delta-S < .02 are required for convergence - some complex models may randomly have an early pass with a rough mesh dip below .02, but the next pass would be above .02 if it ran.

Hi katko,

I followed your advices. I´ve set the convergetion condition to 3 converged passes, and I tried higher frequencies and a smaller mesh, but the results are not valid.

Do you think there is something wrong with the placement of the wave port?

I think the problem is actually due to not enough segmentation/meshing step. The coil turns are quite close each to other and it might be required to assure that surface under the coil can realy represent differece between currents at each turn separately. I am trtying to simulate this with MoM code, but for fine mesh it requires quite a long time. Did you achieve any progress with HFSS so far?

I tried more thinks in HFSS, includying a coaxial waveguide, adding balls to connect the wires..., and I´ve also tried CST creating the helix with unidimensional wires. In CST I got good results in the middle of the simultation, but at the end of that, when the transient in completed there are lots of peaks and the main peaks are diffused.

I still fighting with that, but I think I´m trying to simulate a very difficult problem for a simulation program.

For your comments about the current flow I can see that you understand how antenna works, I´ll really appreciate your help or advices on that.

I´ll also tried to separate the loops. This should only be important in the power radiated, not in the position of the radiation peaks.