Fitting with Convolution of functions

[Vickie Lynch]

When fitting with convolutions, the parameters in the first function of the convolution are not changed during the optimization. Here is the function before and after fitting (not a good fit):

91  name=LinearBackground,A0=0.0,A1=0.0;(composite=Convolution;name=Gaussian,Height=3582,PeakCentre=0.16,Sigma=0.02;name=BackToBackExponential,I=3582,A=100.0,B=100.0,X0=0.16,S=0.02)
91  name=LinearBackground,A0=656.339,A1=0.277;(composite=Convolution;name=Gaussian,Height=3582,PeakCentre=0.16,Sigma=0.02;name=BackToBackExponential,I=-640.883,A=323.506,B=0.00579332,X0=0.072726,S=-4.6928e-07)

This also happens when fitting from the MantidPlot window Fit Function.

There is no wiki page about Convolution.

[Vickie Lynch]

Hi Roman, Do you have time to fix this fitting problem? I need this fixed for my ticket #6940 to have all the required fitting options work correctly.

[Roman Tolchenov]

Added an attribute to unfix the resolution in a fit. Re #7369

Changeset: d135d02c1a06b85d44f2c6852c6cba8bfdb663f7

[Roman Tolchenov]

Fixing the builds. Re #7369.

Changeset: fbb890735592e1738e468b78db6859c47a97255f

[Roman Tolchenov]

Branch

feature/7369_convolution_fit

[Vickie Lynch]

Thanks Roman for adding this so quickly!

[Stuart Campbell]

This ticket has been transferred to github issue ​8215