D:\Metacept\Neurophos\Python_Test_Scripts\Free_Space_Optics\Lumenairy\validation\repro_traced_carrier_121\_d121_common.py:78: UserWarning: D:\Metacept\Neurophos\Python_Test_Scripts\Free_Space_Optics\Reverse_Symmetric_ASM\tx4designstudy121\20260707 dll Tx02-MSOP16.zmx: Zemax surface 9 is a DGRATING.  Its GEOMETRY is imported as the flat/conic surface it is, and its grating data (period 113.7656 um, design order -4, azimuth 0.0 deg) is attached at prescription['diffractives'] -- but NO diffraction is applied by apply_real_lens / apply_real_lens_traced, which propagate the surface as flat.  Run the orders explicitly: put {'doe': rx['diffractives'][k]} in propagate_traced_carrier_chain's groups list (its gap_before 51.5393 mm / gap_after 0.0000 mm are then bookkept for you -- give the group that FOLLOWS the last DOE entry gap_before=0, since gap_after already carries that leg), one order per congruence via propagate_traced_carrier_chain_multi.
  rx = la.load_zemax_zmx(ZMX)
D:\Metacept\Neurophos\Python_Test_Scripts\Free_Space_Optics\Lumenairy\validation\repro_traced_carrier_121\_d121_common.py:78: UserWarning: D:\Metacept\Neurophos\Python_Test_Scripts\Free_Space_Optics\Reverse_Symmetric_ASM\tx4designstudy121\20260707 dll Tx02-MSOP16.zmx: Zemax surface 11 is a DGRATING.  Its GEOMETRY is imported as the flat/conic surface it is, and its grating data (period 113.7656 um, design order -2, azimuth 270.0 deg) is attached at prescription['diffractives'] -- but NO diffraction is applied by apply_real_lens / apply_real_lens_traced, which propagate the surface as flat.  Run the orders explicitly: put {'doe': rx['diffractives'][k]} in propagate_traced_carrier_chain's groups list (its gap_before 0.0000 mm / gap_after 7.0000 mm are then bookkept for you -- give the group that FOLLOWS the last DOE entry gap_before=0, since gap_after already carries that leg), one order per congruence via propagate_traced_carrier_chain_multi.
  rx = la.load_zemax_zmx(ZMX)
=== approx ablation, order (-4, -2), sets ['b', 'c'] ===
   lib C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py  sha256 957f00129f8b467c
   lib C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py  sha256 2d30f1ed7beb3c7e
grid: RN=1024 dx=51.2334 um RS=4 NFC=12288 WF=4.0 NOUT=192 DXO=0.100 um
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.0515, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 12.71 um but the grid has dx = 51.23 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 12.71 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 17575/65536 pixels (26.8%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 17575/65536 pixels (26.8%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 15136/65536 pixels (23.1%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 15136/65536 pixels (23.1%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.0867, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 7.55 um but the grid has dx = 51.23 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 7.55 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 3527/65536 pixels (5.4%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 3527/65536 pixels (5.4%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.1672, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 3.92 um but the grid has dx = 44.91 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 3.92 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 1907/65536 pixels (2.9%) did not converge to tol=4.491e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 1907/65536 pixels (2.9%) did not converge to tol=4.491e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.3606, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.82 um but the grid has dx = 38.43 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.82 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 24634/65536 pixels (37.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 24634/65536 pixels (37.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:4455: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.5393, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.21 um but the grid has dx = 1.51 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.21 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 7944/19600 pixels (40.5%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 7944/19600 pixels (40.5%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
SAMPLING (baseline last group): na_par 0.4049  na_measured 0.5393  na_grid_nyquist 0.4343  exit power above nyquist 7.513e-04
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 17575/65536 pixels (26.8%) did not converge to tol=5.123e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 17575/65536 pixels (26.8%) did not converge to tol=5.123e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 15136/65536 pixels (23.1%) did not converge to tol=5.123e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 15136/65536 pixels (23.1%) did not converge to tol=5.123e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 3527/65536 pixels (5.4%) did not converge to tol=5.123e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 3527/65536 pixels (5.4%) did not converge to tol=5.123e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 1907/65536 pixels (2.9%) did not converge to tol=4.491e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 1907/65536 pixels (2.9%) did not converge to tol=4.491e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 24634/65536 pixels (37.6%) did not converge to tol=3.843e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 24634/65536 pixels (37.6%) did not converge to tol=3.843e-07 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 7942/19600 pixels (40.5%) did not converge to tol=1.508e-08 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 7942/19600 pixels (40.5%) did not converge to tol=1.508e-08 m within 60 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
  done: newton_max_iters 12 -> 60               
  done: newton_poly_order 6 -> 10               
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 3351/65536 pixels (5.1%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 3351/65536 pixels (5.1%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 1906/65536 pixels (2.9%) did not converge to tol=4.491e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 1906/65536 pixels (2.9%) did not converge to tol=4.491e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.3610, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.81 um but the grid has dx = 38.43 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.81 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 24031/65536 pixels (36.7%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 24031/65536 pixels (36.7%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 8872/19600 pixels (45.3%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 8872/19600 pixels (45.3%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
  done: decentred_fit_poly_order 10 -> 14       
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5072: RuntimeWarning: apply_real_lens_traced Newton inversion: 17575/65536 pixels (26.8%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5072: RuntimeWarning: apply_real_lens_traced Newton inversion: 15136/65536 pixels (23.1%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5072: RuntimeWarning: apply_real_lens_traced Newton inversion: 65536/65536 pixels (100.0%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 65536/65536 pixels (100.0%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: amplitude_model='ray_density' energy self-check FAILED -- exit power / aperture-transmitted input power = 0.000000, outside the documented band [0.8900, 1.0500] for ray_subsample=4.  The ray-tube amplitude is energy-conserving only in the GEOMETRIC LIMIT; at finite ray_subsample it loses about 1% at the shipped ray_subsample=8 (design-battery envelope 0.9535-0.9920, converging to 0.9569-1.0000 at ray_subsample=1), and this band is set clear of that whole envelope.  A ratio this far off usually means something else: a fold caustic capping |det J| (see the fold-caustic warning -- use apply_real_lens_gbd / apply_real_lens_fga there), a ray map running off the grid, or an aperture_diameter wider than the traced pupil.  Lower ray_subsample to check convergence.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.4099, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.60 um but the grid has dx = 44.91 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.60 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.5638, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.16 um but the grid has dx = 38.43 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.16 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5072: RuntimeWarning: apply_real_lens_traced Newton inversion: 1735/65536 pixels (2.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 1735/65536 pixels (2.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
  done: newton_fit polynomial -> spline         
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.0887, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 7.38 um but the grid has dx = 51.23 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 7.38 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.1945, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 3.37 um but the grid has dx = 44.91 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 3.37 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.3579, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.83 um but the grid has dx = 38.43 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.83 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
  done: amplitude_model ray_density -> screen   
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.3777, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.73 um but the grid has dx = 38.43 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.73 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 24631/65536 pixels (37.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 24631/65536 pixels (37.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:4455: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.5396, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.21 um but the grid has dx = 1.51 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.21 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 7941/19600 pixels (40.5%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 7941/19600 pixels (40.5%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
  done: remap_sampling full -> lattice          
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 3533/65536 pixels (5.4%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 3533/65536 pixels (5.4%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 24627/65536 pixels (37.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 24627/65536 pixels (37.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 7435/19600 pixels (37.9%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 7435/19600 pixels (37.9%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
  done: fit_radius_beam_factor -> 3.0           
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 70300/262144 pixels (26.8%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 70300/262144 pixels (26.8%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 60094/262144 pixels (22.9%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 60094/262144 pixels (22.9%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.0867, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 7.56 um but the grid has dx = 51.23 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 7.56 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 14215/262144 pixels (5.4%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 14215/262144 pixels (5.4%) did not converge to tol=5.123e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.1675, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 3.91 um but the grid has dx = 44.91 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 3.91 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 7925/262144 pixels (3.0%) did not converge to tol=4.491e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 7925/262144 pixels (3.0%) did not converge to tol=4.491e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:5827: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.3607, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.82 um but the grid has dx = 38.43 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.82 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 98533/262144 pixels (37.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 98533/262144 pixels (37.6%) did not converge to tol=3.843e-07 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\propagators\carrier.py:4455: RuntimeWarning: apply_real_lens_traced: the exit beam converges at NA_exit=0.5395, so the exit wavefront needs dx <= lambda/(2*NA_exit) = 1.21 um but the grid has dx = 1.51 um.  The beyond-Nyquist annulus of the exit phase ALIASES: far-halo energy lands at wrong radii, so r^2-weighted spot metrics (r2m / second moments) read low while EE50/EE80 stay plausible.  Use a finer grid (dx <= 1.21 um) for halo-faithful results, or pass on_undersample="silent" to suppress.
  E_exit = apply_real_lens_traced(
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5067: RuntimeWarning: apply_real_lens_traced Newton inversion: 31720/78400 pixels (40.5%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  return _invert_newton(Xw, Yw, sub_progress=sub_progress)
C:\Users\Tesla\AppData\Local\Temp\claude\C--Users-Tesla\372a2d1f-acbe-4b57-a148-eeae3fe1d729\scratchpad\pin_d2e60ca\lumenairy\elements\_lens_traced.py:5153: RuntimeWarning: apply_real_lens_traced Newton inversion: 31720/78400 pixels (40.5%) did not converge to tol=1.508e-08 m within 12 iterations.  Affected pixels keep their last Newton value, which may carry residual error.  Increase newton_max_iters if this matters for your tolerance budget.
  opl_f, xe_g, ye_g = _invert_newton(Xg, Yg, _want_entrance=True)
  done: ray_subsample 4 -> 2                    

variant                               EE3%    EE6%   EE12%   Ptile%      relL2  dphi_rms      s
-----------------------------------------------------------------------------------------------
BASELINE (shipping defaults)        72.501  92.278  95.636  95.6482  0.000e+00  0.00e+00    384
newton_max_iters 12 -> 60           72.501  92.278  95.636  95.6482  0.000e+00  1.06e-17    407
newton_poly_order 6 -> 10           72.501  92.278  95.636  95.6482  0.000e+00  1.06e-17    425
decentred_fit_poly_order 10 -> 14   72.496  91.950  95.620  95.6351  2.540e-02  1.57e-02    566
newton_fit polynomial -> spline     FAILED: ValueError: _fine_trace_group_exit: the EXACT high-NA final leg needs an axis-centred retr
amplitude_model ray_density -> screen  FAILED: RuntimeError: _fine_trace_group_exit: the EXACT high-NA final leg cannot be sampled for th
remap_sampling full -> lattice      54.768  73.359  75.807  75.8157  3.611e-01  3.96e-01    590
fit_radius_beam_factor -> 3.0       73.158  92.427  95.568  95.5812  2.549e-02  1.59e-02    595
ray_subsample 4 -> 2                72.877  92.801  96.152  96.1643  3.870e-03  1.99e-03    604

dEE3 (points) vs baseline:
  newton_max_iters 12 -> 60                 +0.0000  (dEE6  +0.0000, dPtile  +0.00000)
  newton_poly_order 6 -> 10                 +0.0000  (dEE6  +0.0000, dPtile  +0.00000)
  decentred_fit_poly_order 10 -> 14         -0.0051  (dEE6  -0.3278, dPtile  -0.01313)
  remap_sampling full -> lattice           -17.7325  (dEE6 -18.9189, dPtile -19.83252)
  fit_radius_beam_factor -> 3.0             +0.6574  (dEE6  +0.1491, dPtile  -0.06696)
  ray_subsample 4 -> 2                      +0.3764  (dEE6  +0.5225, dPtile  +0.51610)
