Meep integration: cell size not correctly calculated

[HelgeGehring]

The size of the Mode monitor is not correctly taken into account for bends:

from gdsfactory.components import bend_euler
from gdsfactory.simulation.gmeep import write_sparameters_meep
from gdsfactory.simulation.effective_permittivity import calculate_effective_permittivity

bend = bend_euler(radius=5)
material_name_to_meep = {
    'si': float(calculate_effective_permittivity(3.477 ** 2, 1.444 ** 2, 1.444 ** 2, .22, 1.55, 'te')[0] ** .5)}
write_sparameters_meep(bend, resolution=20, is_3d=False, animate=True, material_name_to_meep=material_name_to_meep)

[simbilod]

The band-aid solution to this is to manually pad the component before simulating it:

    bend = bend_euler(radius=5)
    p = 3
    bend = gf.add_padding_container(bend, default=0, right=p)
    material_name_to_meep = {
        'si': float(calculate_effective_permittivity(3.477 ** 2, 1.444 ** 2, 1.444 ** 2, .22, 1.55, 'te')[0] ** .5)}
    write_sparameters_meep(bend, resolution=20, is_3d=False, animate=True, material_name_to_meep=material_name_to_meep)

We could automate this padding in gmeep.get_simulation conditionned on component size, port location, and port sizes if that is a persistent issue. EDIT: working on it now

[joamatab]

yes, we could pass some padding, or you can use the bend90 function that we have in the docs

https://gdsfactory.github.io/gdsfactory/notebooks/plugins/meep/001_meep_sparameters.html#Port-symmetries

df = gm.write_sparameters_meep_1x1_bend90(c, run=False)

[simbilod]

See PR #512

[joamatab]

Hi Simon

i'd rather set the simulation manually, as we do now

the current solution:

• minimizes extra space
• lets you control the margin manually

write_sparameters_meep_mpi_1x1_bend90 = gf.partial(
    write_sparameters_meep_mpi,
    ymargin_bot=3,
    ymargin=0,
    xmargin_right=3,
    port_symmetries=port_symmetries.port_symmetries_1x1,
)

what do you think?

[simbilod]

Agreed

[joamatab]

another way to automatically fix these issues is to add port_markers

[joamatab]

we could do this as a FdtdSimulator class that can have all the FDTD as backends

• meep
• lumerical
• tidy3d

we could also add 3D rendering capabilities

what do you think?

@HelgeGehring
@flaport

[HelgeGehring]

another way to automatically fix these issues is to add port_markers

seems like the best approach!

[simbilod]

This is a good opportunity to allow these port markers to hold height and mode information to allow multilayer and multimode simulations #502

[joamatab]

see function to add port markers in the docs