SCF convergence

Message

neeti · #1 Post by **neeti** » Tue Aug 01, 2006 3:20 pm

For the DOS of a coordination complex(NELECT, ISMEAR, ISPIN, IDIPOL specified), SCF cycle shows following disruption before continuing the process;

LDA part: xc-table for Pade appr. of Perdew
POSCAR, INCAR and KPOINTS ok, starting setup
FFT: planning ... 1
reading WAVECAR
entering main loop
N E dE d eps ncg rms rms(c)
RMM: 1 0.551079708322E+04 0.55108E+04 -0.59164E+04 544 0.560E+02
.........................................
.........................................
RMM: 11 -0.438556491891E+03 -0.90660E+00 -0.46151E+00 1583 0.235E+00
RMM: 12 -0.438661380328E+03 -0.10489E+00 -0.60059E-01 1689 0.631E-01BRMIX: very serious problems
the old and the new charge density differ
old charge density: 346.00000 new 327.00001
0.713E+01
RMM: 13 -0.350349984010E+03 0.88311E+02 -0.13020E+03 1650 0.515E+01 0.568E+01
........................................................................
.........................................................................

This SCF cycle does not converge even after 300 steps. Does the message above regarding BRMIX imply that some paramter needs to be altered? Any suggestions?

#2 Post by **admin** » Wed Aug 02, 2006 1:07 pm

a lack of 19 electrons really is a serious problem, how much does NELECT actually differ from the valence charges according to the pseudopotentials you use? the difference of the charges given in that message should be equal to this difference, in the first iteration in which the charges really are mixed (NELMDL). I don't suppose you want to calculate a +19-charged cluster (!)

neeti · #3 Post by **neeti** » Thu Aug 03, 2006 3:53 pm

according to the valence charges in pseudopotentials the total number of electrons should be 332, but NELECT is set equal to 328 in INCAR. Should I be changing the mixing parameters when mentioning NELECT?
<span class='smallblacktext'>[ Edited Thu Aug 03 2006, 05:57PM ]</span>

#4 Post by **admin** » Thu Aug 10, 2006 9:44 am

according to the POSCAR and POTCAR you use, the total valence charge is in fact 346, not 332, please check your input