[Wien] projected DOS problem

[Andrea Salguero]

Dear all
I have a triclinic system, a organic compound, with 28 atoms in the unit cell 
but only one copper atom. The scf cycle has converged very well (ENE, DIS 
criteria 0.0001), I have calculated DOS and bandstructure without any 
problem. I wanted to project the Cu-d states in other base, so  I used  one 
of the default matrices .cf* stored in SRC_templates for d-states 
(case.cf_d_eg_t2g) and I did the following:

-> in  case.inq file I introduced the Fermi energy, latom and jatom, this file 
looks like:
FULL        /FULL      ( j=1/2 or 1/2,-1/2 and 1/2,1/2 extra)
DOSYM       /NOSYM     ( symmetrization is/is not performed )
-2.  2.                 ( energy window)
0.26577                     ( fermi energy)
1                      ( number of atoms for which the DOS are calculated)
1  2                    ( jatom  latom   (index of ineq. atom, orbital quantum 
number - only one peratom)

-> ./qtlpara qtl.def
this generates a file .qtl1 and I introduce that in tetra.def
->modify .int to generate the .dos files 
-> ./tetra tetra.def
 -> and then plot

everything ends without any error message, but when I compared the DOS in the 
new base with the one calculated previously (before to change the basis), all 
is the same except the height of the peaks, the heights are lower than the 
ones calculated in the old base. Then I superposed old and new DOS graphics 
and I obtain in the DOS with the new base one half of the peaks compared to 
the first case, only when I multiply by 2 I obtain the same heights. I tried 
other matrices given in SRC_templates and I obtain again the half.

I tried the same procedure, with the same matrix in a cubic system and in that 
kind of system this problem didn't appear. Could somebody give me some help? 

Best Regards
Andrea