Reply To: Loss Due to Metal Electrodes Sanity Check

[Guilin Sun]

It is a good practice to do convergence testing by gradually increasing simulation spans. Since mode profile has long tails which can penetrate into the metal electrodes, you may need insrease y span longger. Of course the tail may not be exactly zero. So you can set the colorbar to certain value, say 1e-8 or smaller (intensity, eg, the tail min intensity at the ymax), I believe the loss will not change any further in most cases. You can compare the result with two different y span and see how the loss changes.