[Wien] Sum of PDOSes doesn't give total DOS

[David Holec]

Dear Wien2k users & developers,

I'm getting a bit puzzled by following problem: when I calculate total 
DOS of a system using tetra, and at the same time calculate PDOS of that 
system, the sum of DOSes does not add up to total DOS.

Let me be a bit more specific: I have a 4 atomic unit cell of AlN, 2 
atoms being Al and two N. My assumption would be that when I use 
following int-file

Title
  -1.50 0.002 2.500 0.003   EMIN, DE, EMAX, Gauss-broadening(>de)
     11                     NUMBER OF DOS-CASES specified below
     0    1   total         atom, case=column in qtl-header, label
     1    1   Atom1 tot
     1    2   Atom1 s
     1    3   Atom1 p
     1    4   Atom1 pz
     1    5   Atom1 px+py
     2    1   Atom1 tot
     2    2   Atom1 s
     2    3   Atom1 p
     2    4   Atom1 pz
     2    5   Atom1 px+py

then it would hold e.g.:

   Atom1 tot + Atom2 tot = total

or:

   Atom1 s + Atom1 p = Atom1 tot

When I integrate the total DOS (column 2 of the output corresponding to 
the above tetra input) over -20eV to 0eV, I got 16.004e which is very 
close to 16e I am expecting in the in this region (I am basically 
missing 4*2 1s electrons which are deeper than my energy window). But 
the same integral for Atom1 tot (column 3 in the output) gives 1.4285e 
whereas for Atom2 tot (column 8 in the output) it gives 8.5117e. Their 
sum is 9.9404e...

I assume it is tricky to try to distinguish which electron belongs to 
which atom in solid state but still I would expect that the sum of these 
two would be give the total DOS.

Could any help me either how to get out of this (can I simply 
renormalise the PDOSes in order to get correct number of electrons?) or 
tell me where my thinking goes wrong?

Many thanks in advance,
David