Electronic Loops of Band structure calculation

[zhichengzhong]

Dear Administrator:

When I do a standard band structure calculation (set ICHARG =11). In the OSZICAR, there is loops similar to normal self-consistent calculation:
N E dE d eps ncg rms rms(c)
DAV: 1 0.452735366526E+04 0.45274E+04 -0.11916E+05 26060 0.210E+03
DAV: 2 0.102403969618E+04 -0.35033E+04 -0.32526E+04 23616 0.608E+02
DAV: 3 0.270812852845E+02 -0.99696E+03 -0.90601E+03 26568 0.277E+02
DAV: 4 -0.264430771361E+03 -0.29151E+03 -0.26423E+03 35964 0.143E+02
DAV: 5 -0.325544266977E+03 -0.61113E+02 -0.57239E+02 43136 0.696E+01
DAV: 6 -0.331464263806E+03 -0.59200E+01 -0.58051E+01 43136 0.273E+01
DAV: 7 -0.332204182696E+03 -0.73992E+00 -0.73612E+00 43088 0.104E+01
..............

And CHGCAR file also have a very small change.

Now, the problem is
1: For ICHARG =11, that's means the charge density remains constant in the run, why does the CHGCAR have tiny change.

2: Band structure calculation is non-self consistent, and all the k points is treated independently. So I don't understand what's the meaning of these loops. What is updated with the loops?

Thanks

[admin]

1) CHGCAR should NOT be updated if ICHARG=11
2) there have to be a few electronic steps to converge the wavefunction and eigenvalues to the scf solution for the new set of k-points