Why do solution frequency affect the simulation result in HFSS?
First, sorry about my poor english.
Second, I have checked many posts and books but still cant know where is wrong.
I place a resonator inside a waveguide and wanna know resonant frequency.
If i change the solution frequency(sol feq), the resonance frequency(res feq) change.
The difference is too large for me, so that's the reason why i'm here asking for help.
For example, i set two different sol feq, 2.7GHz and 2.9GHz, and result is resonance at 2.966GHz and 2.979GHz.
I try more strick adaptive conditions but the res feq do not come close and just stop at 3.008GHz and 3.035GHz.
I am really confused about it. Maybe there's some very trivial things i got mistake.
Is this because of the accuracy limit from HFSS itself?
Or if i wanna do a very accuracy simulation, there's some option i should aware?
Or different sol feq just always gives us a different result?(but why?)
Most of people seems like just set the solution frequency at "interested" frequency,
and if more accurated result is needed, just set sol feq more close to the res feq.
In my case, set sol feq too close to the res feq will get very strange result(resonance disapear) or just diverge.
I use fast sweep and always set sol feq at the center of sweep range.
Also tried discrete sweep but the problem still the same.
Attachment is the HFSS file of this structure.
Hope there's some suggestion or something.
Regards
It sounds like it's probably convergence. There is not a fine enough mesh to accurately determine the resonance frequency, such that a change in solution frequency changes the solution. This can also happen if you converge with adaptive meshing at a frequency which supports a different mode -- the resulting mesh will be adequate for the second mode, but not the original one you were looking for.
Increase the convergence criteria (number of passes, delta S), and you should see better results. Keep the convergence frequency close to the expected resonance frequency, and always do discrete sweeps for best results.
When you define the solution frequency then HFSS generates mesh according to that solution frequency to solve the problem so it directly affects the accuracy. Just keep this thing in the mind that solution frequency define the mesh so it is in reality mesh which controls accuracy of your problem.
One way to check this thing is that define mesh for your problem manually rather relying on HFSS generated mesh and then check if still solution frequency is affecting your results in the same way. As it is the mesh definition which dictates the accuracy.
What seems that your solution is not converging properly
to account for the worse performance set the solution frequency to the largest frequency in your bandwidth , instead of the center frequency. and increase the number of passes