-
|
Hi all, sorry, newbie here, again, I'm stuck:( any help or tips would be highly appreciated. I've now to geometry and meshing. I've also got a lumped port and excitation, but I am running into the following when creating the NF2FF box (I think this is the last piece I need to actually see any results .. I hope so;): what is "not enough lines in some direction" supposed to mean? what direction? fwiw, here is my code for port and excitation and such: def _setup_excitation(self):
helix_via_shell_width = 1.0 / 2 - 0.3 / 2
helix_via_radius = 0.3 / 2 + helix_via_shell_width / 2
self._excitation_freq = (self._max_freq + self._min_freq) + self._min_freq
self._excitation_bw = (self._max_freq + self._min_freq) / 2
self._fdtd.SetGaussExcite(self._excitation_freq, self._excitation_bw)
self._fdtd.SetBoundaryCond(['MUR', 'MUR', 'MUR', 'MUR', 'MUR', 'PML_8'])
feed_x0 = self._helix1_waypoints[0][0]
feed_y0 = self._helix1_waypoints[0][1] - self._helix_thickness / 2
feed_start = [feed_x0, feed_y0 - helix_via_radius, -self._helix_thickness]
feed_stop = [feed_x0, feed_y0 - helix_via_radius, self._helix_thickness]
feed_resistance = 50
# apply the excitation & resist as a current source
self._helix1_port = self._fdtd.AddLumpedPort(1, feed_resistance, feed_start, feed_stop, 'z', 1.0, priority=5,
edges2grid='xy')
# do I need that? What does it do anyway?
# self._mesh.SmoothMeshLines('all', mesh_res, 1.4)
# add the NF2FF recording box
self._nf2ff = self._fdtd.CreateNF2FFBox()this is how the model looks in the viewer:
|
Beta Was this translation helpful? Give feedback.
Replies: 3 comments
-
|
fwiw, I've tried to "smooth the mesh" self._mesh_res_farfield = round(self._max_freq_lambda / 30)
self._mesh_res_nearfield = round(self._mesh_res_farfield / 2)
self._mesh_res_antenna = round(self._mesh_res_nearfield / 5, 1)
self._mesh_res_antenna_turn = round(self._mesh_res_antenna / 2, 1)
// snap
# do I need that? What does it do anyway?
self._mesh.SmoothMeshLines('all', self._mesh_res_farfield, 1.4)even though I don't understand what exactly it does, it does add more mesh lines
however, the traceback persists :( Exception: Error::CreateNF2FFBox: not enough lines in some direction here are the orthogonal projections of the mesh
|
Beta Was this translation helpful? Give feedback.
-
|
Ok, actually, I could brute-force it by smoothing the mesh to the finest resolution at which I am creating mesh lines: # this does NOT work!
# self._mesh.SmoothMeshLines('all', self._mesh_res_farfield, 1.4)
# this does NOT work!
# self._mesh.SmoothMeshLines('all', self._mesh_res_nearfield, 1.4)
# this does NOT work!
# self._mesh.SmoothMeshLines('all', self._mesh_res_antenna, 1.4)
# this DOES indeed work!
self._mesh.SmoothMeshLines('all', self._mesh_res_antenna_turn, 1.4)
# add the NF2FF recording box
self._nf2ff = self._fdtd.CreateNF2FFBox()
now, above seems pretty excessive, and I am now even more confused:( sure, "it does work" as in it doesn't traceback .. however currently, my conclusion would be: openEMS can only use uniform cells, and the number of cells is thus determined and driven by the finest feature in the geometry. is that true? the "finest feature" in my current geometry are the Vias (which yes, I don't need that at 1GHz, but I was too lazy up til now to remove this again after I added it in the first place):
|
Beta Was this translation helpful? Give feedback.
-
|
Alright, I figured it out by bisection of geometry vs the working (excessive) mesh setup with smoothing - turns out, I was missing sufficient meshing in z-axis only So I got geometry and meshing, but I guess my port/excitation is still wrong. Or there is another problem I am totally unaware of. Anyways, different issue, this one seems fixed. fwiw, I also simplified the "vias" (removed the hollow shape, now vias are solid). The antenna segments are still "round" wires, not flat, but that shouldn't make a fundamental difference .. hopefully .. at least it runs fast enough: The simulation now runs =(
|
Beta Was this translation helpful? Give feedback.
Alright, I figured it out by bisection of geometry vs the working (excessive) mesh setup with smoothing - turns out, I was missing sufficient meshing in z-axis only
So I got geometry and meshing, but I guess my port/excitation is still wrong. Or there is another problem I am totally unaware of. Anyways, different issue, this one seems fixed.
fwiw, I also simplified the "vias" (removed the hollow shape, now vias are solid). The antenna segments are still "round" wires, not flat, but that shouldn't make a fundamental difference .. hopefully .. at least it runs fast enough: