I know there as been a lot discussion going on with nonlinear fitting. Still I have some problems with this. I am doing an automated fit routine where I fit Lorentzians to a huge dataset. I get the initial parameters from an elaborate automated data analysis. Although the initial parameters are very close to the final fit parameters (probably only a factor of 2 off for area and width and 1% for the center) only every once in a while my fit converges. I am aware that Lorentzians are already sofisticated to fit automatically but since my initial parameters are so close I do not see where is goes wrong. By the way, I have the weight unwired (I do not understand completely what this does) and I use 1e-40 and 1000 as termination parameters. The LabView version I use is 8.5.1 and I used the Lev_Mar FormulaString nonlinear fit Vi. Thanks in advance, best wishes, Martin.
