hi
itaifrenkel posted a message below. it would help:
HFSS using convergence in order to get an accurate solution. After creating an initial mesh it starts doing adaptive passes until the desired error residual is achieved. In each stage it refines the mesh and resolves the problem. The resulting S-Matrix is compared to the previous one. The maximum delta is computed in percentage. If the max error found is 3% and you specified Maximum delta S per pass.to be 0.02=2% then it will continue with another meshing and solving. If the max error found is reduced to 1.5% the solution is "converged" and the simulation will stop.
In order to control this process you have a few options:
1) Maximum number of passes - Allows you to stop after several passes. Even if you want faster solution times I recommend not to use this option and set it to 20.
2) Maximum delta S per pass - Allows you to define the desired accuracy for the solution. For example, in many Signal Integrity applications 4% is more than enough.
3) Refinment per pass - for simple problems increase from 20% to 30%. This controls the amount of mesh refignment between passes.
4) Minimum converged passes - I recommend to set it to 1. This allows you to force HFSS to continue more adaptive passes even after it converged.
5) Port accuracy. The default is 2%. This controls the accuracy of the 2D solution on the port. If you are solving ports only (for example to calculate effective lambda for a transmition line) - I recommend to increase accuracy from 2% to 0.3% or even 0.1%.
Itai