XDATCAR changes into unphysical values

[hosfa]

I am doing a simple ionic relaxation for SiC.
my POSCAR has somehow exact positions.
for INCAR:

ISMEAR = -1
SIGMA = 0.001
ISPIN = 1
ENCUT = 420.00 eV
ENAUG = 840.00 eV
NELM = 40
EDIFF = 1E-05
NSW = 10
ADDGRID = .TRUE.
IBRION = 1
ISIF = 3
EDIFFG = -1E-03


but after each ionic step I receive this error:

entering main loop
N E dE d eps ncg rms rms(c)
DAV: 1 0.129828706002E+04 0.12983E+04 -0.39023E+04 17072 0.226E+03
DAV: 2 0.244602139079E+03 -0.10537E+04 -0.99785E+03 11096 0.856E+02
DAV: 3 -0.328803301434E+02 -0.27748E+03 -0.27512E+03 13144 0.249E+02
DAV: 4 -0.586889148843E+02 -0.25809E+02 -0.25784E+02 13584 0.921E+01
DAV: 5 -0.604665811512E+02 -0.17777E+01 -0.17775E+01 14344 0.309E+01 0.157E+01
DAV: 6 -0.533393637525E+02 0.71272E+01 -0.77105E+00 16728 0.191E+01 0.583E+00
DAV: 7 -0.532262676304E+02 0.11310E+00 -0.14832E+00 14320 0.425E+00 0.326E+00
DAV: 8 -0.532189642026E+02 0.73034E-02 -0.11581E-01 15696 0.288E+00 0.554E-01
DAV: 9 -0.532109608027E+02 0.80034E-02 -0.12442E-02 15624 0.127E+00 0.921E-02
DAV: 10 -0.532115058272E+02 -0.54502E-03 -0.22443E-03 15320 0.215E-01 0.847E-02
DAV: 11 -0.532115310038E+02 -0.25177E-04 -0.21751E-04 12624 0.993E-02 0.124E-02
DAV: 12 -0.532115340379E+02 -0.30341E-05 -0.17904E-05 11056 0.301E-02
1 F= -.53211534E+02 E0= -.53211534E+02 d E =-.532115E+02
BRION: g(F)= 0.682E-03 g(S)= 0.200E+03
bond charge predicted
N E dE d eps ncg rms rms(c)
ERROR FEXCP: supplied Exchange-correletion table
is too small, maximal index : 15639


In the XDATCAR I noticed this strange values:(by strange I mean the lattice vectors has become 1/3!!!!)
PBE calculation of SiC primitive c
1
3.093313 0.000000 0.000000
-1.546657 2.678888 0.000000
0.000000 0.000000 10.131661
Si C
4 4
PBE calculation of SiC primitive c
1
1.132618 0.000000 0.000000
-0.566309 0.980876 0.000000
0.000000 0.000000 3.693102
Si C
4 4
Direct configuration= 1
0.00000000 0.00000000 0.00007817
0.33333333 0.66666667 0.24982966
0.66666667 0.33333333 0.50007817
0.33333333 0.66666667 0.74982966
0.00000000 0.00000000 0.18802582
0.33333333 0.66666667 0.43741162
0.66666667 0.33333333 0.68802582
0.33333333 0.66666667 0.93741162

[admin]

Note that after 10 ionic steps (NSW=10) the system could be far away from thre relaxed structure.

[hosfa]

[quote="10) the system could be far away from thre relaxed structure.[/quote"]
Thanks!
Considering the fact that my POSCAR is quite accurate It shouldn't change that much!
Any way, I cannot find any solution for the error!