search

ATTPC / Spyral

A Python analysis library for AT-TPC data
GNU General Public License v3.0
2 stars 2 forks source link

Distorition in cross section #199

Closed gwm17 closed 2 months ago

gwm17 commented 2 months ago

Lots of users are reporting shoulders in low center-of-mass angle elastic scattering cross sections. Personally, have not been able to replicate, but it is possible that it is due to some configuration of the parameters or even using one of the solvers in particular. Need to verify against previous analysis frameworks (i.e. ATTPCROOT, IgorPro) and make sure we aren't missing something here.

gwm17 commented 2 months ago

If anyone has examples to show here that would be great!

gwm17 commented 2 months ago

Early word is that it looks related to choosing between L-BFGS-B and leastsq where leastsq seems to have some bad behavior for some datasets...

gwm17 commented 2 months ago

More analysis has yielded interesting results. It looks to be related to the inclusion of error in the fitting.

gwm17 commented 2 months ago

Solved. Fitter with stricter confidence requirements (i.e. LeastSq) will often reject lower energy data, which, in scattering often will correspond to a dip at low CM angles.