Bugfix: Don't use symmetries for weighting potentials #216
Conversation
|
|
||
| if for_weighting_potential && world_Δφ > 0 | ||
| world_φ_int = SSDInterval{T, :closed, :open, :periodic, :periodic}(0, 2π) | ||
| world_Δφ = width(world_φ_int) |
There was a problem hiding this comment.
Do we always want it to be full 2π ? I think this makes the full_2π keyword obsolete?
We could set full_2π = true here...
There was a problem hiding this comment.
Yes, I think it is obsolete. It's nowhere used. I will remove it.
I would keep it as we did it before. Always calculate the weigting potentials for 2π.
Only if it is completly phi-symmetric. Than, also only 2D.
I think we should refactor the symmetry handling, #215, in general. But for now we just need a quick fix.
src/Simulation/Simulation.jl
Outdated
| if full_2π || (for_weighting_potential && (world.intervals[2].left != world.intervals[2].right)) | ||
| L, R, BL, BR = get_boundary_types(world_φ_int) | ||
| int_φ = Interval{L, R, T}(endpoints(world_φ_int)...) | ||
| if full_2π || (for_weighting_potential && (world_φ_int.left != world_φ_int.right)) |
There was a problem hiding this comment.
.. and then we can check only the statement if full_2π here?
|
I am a bit surprised that the tests for the segBEGe detector did not fail ( |
Because contact 1 lied (at least partially) inside the segment of the world which was covered by the symmetry defined for the electric potential. And also the event was spawned inside this volume. So a sum of 2 was produced. However, the pulse shape is probably also wrong. |
Crucial bugfix for the calculation of the weighting potentials.
Bug: The grid for the weighting potential were initialized with the symmetries of the world which are only valid for the
electric potential and not the individual contacts.