Closed aoeftiger closed 4 years ago
PR https://github.com/SixTrack/pysixtrack/pull/41 fixes the rotation of the curvature.
The following notebook documents the implemented scheme (same as above, just re-run with the PR https://github.com/SixTrack/pysixtrack/pull/41): https://github.com/aoeftiger/sixtracklib_pyht_playground/blob/656a725f083dfeabdf60f20e54909884f71ab754/sis100/2020-01-10--multipole-error-comparison-MADX-STL/ThinDipole_with_dpsi-entering_on_axis.ipynb
==> MAD-X produces a 10^{-5} error in the px
and py
kicks (which should be equal to the given k0l
strength of 0.01
):
enter TRACK module
one pass is on
++++++ table: tracksumm
number turn x px
1 0 0.00000000000000e+00 0.00000000000000e+00
1 1 5.00066682337395e-03 1.00003333466672e-02
y py t pt
0.00000000000000e+00 0.00000000000000e+00 0.00000000000000e+00 0.00000000000000e+00
-5.00091686713360e-03 -1.00008333841694e-02 -8.80997659244605e-05 0.00000000000000e+00
s e
0.00000000000000e+00 0.00000000000000e+00
1.00000000000000e+00 0.00000000000000e+00
exit TRACK module
(Note the px
of 1.00003333466672e-02
instead of 1.0e-02, as well as the py
of -1.00008333841694e-02
instead of -1.0e-02)
PySixTrack's application of a transverse rotation misalignment error (
dpsi
in MAD-X) to a thin multipole with a dipole componentk0l
is missing the rotation of the curvature termshxl
andhyl
. (cf. email thread between @rdemaria and me)The problem is documented here: https://github.com/aoeftiger/sixtracklib_pyht_playground/blob/04e7c24cb318785171a4b3b757b2ea6c99c9b8f7/sis100/2020-01-10--multipole-error-comparison-MADX-STL/ThinDipole_with_dpsi-entering_on_axis.ipynb
Important to know: MAD-X Version 5.05.01 seems to yield slightly wrong results for strong
k0l
components -- as documented in the notebook above. This systematic error vanishes quickly for weakerk0l
<< 1e-4.