You will find these kinds of questions answered in http://www.wien2k.at/reg_user/textbooks/DFT_and_LAPW-2_cottenier.pdf , in particular chapter 6.
> I guess what I really asking is: are the energy eigenvalues listed at > the end of case.scf1 (or in case.energy, or in case.vector) NOT used > to calculate the radial wavefunctions in the LAPW basis? Indeed, these are not the ones that are used. Using individual eigenvalues instead of a fixed linearization energy would miss the point of making a linearization. > Secondly, the definition of a local orbital (equation 2.7 in the > WIEN2k manual) describes two energy parameters E_1,l; E_2,l. In the > case.scf1 file there is only one energy listed for the local orbital, > is this E_2,l (and the global LAPW E_l is used as E_1,l)? Yes and no. Yes, this is E_2,l. No, it is not the global linearization energy, but rather the linearization energy for the regular LAPW with that l-value. The latter can be equal to the global one, but need not to be so (see the paragraph underneath eq. 4.4 of the pdf given above). Stefaan