symmetry error.

[nkxirainbow]

Dear developers:
I try to do a scf calculation for rhombohedral CaMnO3.
The "static configuration" gives point symmetry C_3h.
However "dynamic configuration" returns point symmetry C_2h.
I do not know why there are two different symmetry:C_3h and C_2h. If I add SYMPRE=1E-4, everything works well. If I add SYMPRE=1E-5, the above error persist.
The following is my input file and error messages.

INCAR:
System = CaMnO3
ISPIN = 2
MAGMOM = 1*3 1*-3 8*0

POSCAR:
CaMnO3
1.05000000000000
4.0774467077209886 4.0774467077209886 0.0091831936813522
4.0682635140396348 -0.0000000000000004 -4.0682635140396357
-0.0000000000000006 4.0682635140396348 -4.0682635140396357
Mn Ca O
2 2 6
Direct
0.03531404393827 0.01489531868724 0.01489531868724
0.53532453929230 0.51489182023590 0.51489182023590
0.27926945855577 0.26691018048141 0.26691018048141
0.77927223704644 0.76690925431785 0.76690925431785
0.75867416586710 0.27783562163822 0.76565459085646
0.25867615741238 0.77783242870719 0.26565898517321
0.75867416586710 0.76565459085646 0.27783562163822
0.25867615741238 0.26565898517321 0.77783242870719
0.25867615741238 0.77783242870719 0.77783242870719
0.75867416586710 0.27783562163822 0.27783562163822

Errors printed on screen:
VERY BAD NEWS! internal error in subroutine INVGRP:
inverse of rotation matrix was not found (increase SYMPREC) 4



Thank you in advance:)
<span class='smallblacktext'>[ Edited ]</span>

[admin]

if the error vanishes if SYMPRE is increased to 1e-5, the symmetry of the cell is broken if the positions of (one of the) the atoms are read with a higher precision, i.e. some of the positions are in contradiction to the C_3h symmetry if they are read with higher accuracy.

[nkxirainbow]

Dear administrator:
Thanks you very much for your reply.
Could you tell me what is the symmetry difference between:"dynamic configuration"(all velocities are zeros) and "static configuration"?