[Wannier] R: Issue with convergence

[Gibertini Marco]

Dear Carlo,

the attachments are missing, but if you say that the bands do not cross 
you do not have to disentangle them even if their energy ranges overlap 
(just make sure that your path includes all relevant high-symmetry lines).

In addition to what Giovanni said, maybe you could improve the 
wannierization by choosing a better initial position for your s orbital.

As far as the NaN are concerned, it must a problem related to the unk files.
These files are not needed for the wannierization (that's why everything 
goes smoothly) but they are used to plot the Wannier functions in real 
space.
Since they are binary files produced by pw2wannier90 the problem might 
arise from your compilation.
For instance, did you use the same compiler for pw, pw2wannier90 and 
wannier90?

Hope this helps!
Best,
Marco

-- 
****
Marco Gibertini
****************************************
Post-doctoral Research Scientist
Theory and Simulation of Materials
École Polytechnique Fédérale de Lausanne