EFIELD and problem with convergence

[spakinfo]

Dear all,

I'm trying to do electronic structure calculations for a slab under a finite electric field. My system has hexagonal symmetry, and I use the following tags in INCAR to switch on a saw-like electric field of 0.3 eV/A along c axis:

ISMEAR =0
NSW = 1
IBRION = -1
ISPIN = 1
EDIFF = 0.00001
EDIFFG = -0.005
ICHARG = 2
ISYM=0
LMAXMIX=6
ENCUT=500
EFIELD =0.3
LDIPOL=.TRUE.
IDIPOL= 3
NELM=200
PREC=A

With the above settings, however, electronic convergence seems to be out of reach. Without them, on the other hand, there seems to be no such a problem. Now I'm wondering, If this issue is due to
1) The hexagonal symmetry of my structure
2) The amplitude of electric field
3) mixing parameters

Any help is greatly appreciated.
Spak

[admin]

please first of all try to fix the origin of the dipol. Usually the bad electronic convergence is caused by a permanent shift of the origin of the dipolemoment

[spakinfo]

Thanks admin.
Does it matter where to fix the dipole? Should I fix it somewhere inside the slab (e.g. at the center) or elsewhere in the vacuum region?