Problem with multishell model - glitches at high Q

[ajj]

The multishell model has artifacts that appear when the dimension of the core is greater than some value (depends on Q range whether the artefacts are within the range).

Same problem appears in IGOR, so it is an underlying model issue. Suspect a problem with either integration or bessel function precision.

This needs fixing ASAP

[butler]

This probably is best worked through as part of new sasmodels. Since has existed in IGOR code for a while already and will require probably a fair amount of sleuthing that we may not have the bandwidth for we won't hold up the release 3.1 to fix this

[butler]

once again running out of bandwidth to check this for the upcoming release 4.0 — moving to next release

[pkienzle]

This ticket needs specific parameter values to check against, as well as a description of the artefacts. Many of the models have been updated to use more accurate bessel function approximations, at least for small qr, so this may already have been fixed.

[richardh]

Symptom - regular glitches in I(Q) at high Q when using polydispersity in any core shell model.

FIX - users should increase number of steps NPTS and/or number of sigmas NSIGS to integrate over, in the polydispersity integral