[Wien] Problem with jumping forces

[Veerle Vanhoof]

Dear Wien user,

I am having some problems with the force in my calculation.
I'm working in a supercell with a cluster of 2 Co-atoms near substitutional positions in an fcc Ag host (32
atoms in the cell), the calculations are performed for a nonmagnetic case, with spin-orbit-coupling.
When I start my calculation from a converged case (the same happens for an nonconverged case) and I perform two subsequent force-convergence-calculations, the resulting forces on one of the atoms is drastically different, while there is no change for all the other atoms. Other parameters (EFG, ENE,...) don't change:

first calculation:

:FGL006:   6.ATOM                        41.594          0.000        -63.545
:FGL006:   6.ATOM                        41.592          0.000        -63.541
:FGL006:   6.ATOM                        41.596          0.000        -63.549
:FGL006:   6.ATOM                         4.932          0.000          0.345


:FGL007:   7.ATOM                       -17.368         18.270        -24.799
:FGL007:   7.ATOM                       -17.364         18.268        -24.796
:FGL007:   7.ATOM                       -17.357         18.270        -24.798
:FGL007:   7.ATOM                        34.747        -21.567         -9.663

second calculation:

:FGL006:   6.ATOM                        41.579          0.000        -63.545
:FGL006:   6.ATOM                        41.571          0.000        -63.544
:FGL006:   6.ATOM                        41.571          0.000        -63.544
:FGL006:   6.ATOM                         4.936          0.000          0.347

:FGL007:   7.ATOM                       -17.252         18.237        -24.759
:FGL007:   7.ATOM                       -17.150         18.205        -24.723
:FGL007:   7.ATOM                       -17.151         18.205        -24.723
:FGL007:   7.ATOM                       241.417         -0.935         -5.103

This problem is very similar to one I had some time ago with a monoclinic CXZ-lattice. That problem could be solved by using a lower symmetry and was possibly a consequence of a bug for the monoclinic CXZ-lattice. 
This time however, I'm working with a (orthorombic) CXY-lattice. What can be the problem in this case?

In attachment you can find the used struct-file.
Other parameters: RKM=7, 12 k-points in IBZ (4x4x4), GGA (13).
  

Thanks a lot!
Veerle Vanhoof





---------------------------------------------------------------------------
Veerle Vanhoof
  Katholieke Universiteit Leuven
  Instituut voor Kern- en Stralingsfysica
  Celestijnenlaan 200 D
  B-3001 Leuven
  BELGIUM
  Phone: +32 (0)16 32 71 45
  Fax: +32 (0)16 32 79 85
  email: Veerle.Vanhoof at fys.kuleuven.ac.be
---------------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20030709/a38e1561/attachment.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dimernonmagnso.struct
Type: application/octet-stream
Size: 4838 bytes
Desc: not available
Url : http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20030709/a38e1561/dimernonmagnso.obj