[Ifeffit] Contributions from 2 paths

[Matt Newville]

Hi Matt,

>       I am analyzing Neodymium Oxide with G.I.Feffit.  I have two paths and
> I'm having trouble showing the contributions (chi data, Forward and back
> Fouirer Transform) to the fit from each path.  I ran into a problem with the
> Sigma2 for the second path, sigma2_2.

> When I guess sigma2_2 I can print out the individual path contributions to
> the fit with the fitted parameters factored in.  If I define, def ( ),
> sigma2_2 = sigma2_1 then the path contributions that I try to print out do
> not have the debye-waller parameters factored in.

Hmm, it seems like the do to me.   From sigma2_2=sigma2_1.log it says
  Variable
    ...
    sigma2_1   =  0.00293519 +/-  0.03474220
    ...
    sigma2_2   =  0.00293519 := sigma2_1

  Path 1
    ...
    ss2 =     0.002935
 Path 2
    ...
    ss2  =   0.002935

> I have tried the analysis with feffit245 and the path contributions it shows
> seem be consistent with the fit.  So, I figure I must be writing something
> wrong in my command file.

The input files look OK to me.  There is a concern that your results have very
large uncertainties, for  example that e0_1 refines to 13.4 +/- 124.  I'd guess
that starting with a single e0 for the two paths might help.

Or perhaps I'm just not understanding the question???

--Matt