[Wien] false convergence in structure optimization

[Jian ZHOU]

Dear all,

I am doing a structure optimization using the PORT method. Since I only
optimize the positions of atoms automatically, and change the lattice
constants by hand, I find that in some cases(different lattice
constants), the force converge criterion can not be reached.

For example, in one calculation, I got following results:
grep :FGL001 *mini
:FGL001:   1.ATOM     1.461          1.461          0.000
:FGL001:   1.ATOM     0.643          0.643          0.000
:FGL001:   1.ATOM     0.153          0.153          0.000
:FGL001:   1.ATOM     1.145          1.145          0.000
:FGL001:   1.ATOM     1.228          1.228          0.000
:FGL001:   1.ATOM     1.208          1.208          0.000
:FGL001:   1.ATOM     1.284          1.284          0.000
:FGL001:   1.ATOM     1.261          1.261          0.000

grep :FGL002 *mini
:FGL002:   2.ATOM     0.000          2.010          0.000
:FGL002:   2.ATOM     0.000          1.127          0.000
:FGL002:   2.ATOM     0.000          0.689          0.000
:FGL002:   2.ATOM     0.000          1.736          0.000
:FGL002:   2.ATOM     0.000          1.730          0.000
:FGL002:   2.ATOM     0.000          1.634          0.000
:FGL002:   2.ATOM     0.000          1.737          0.000
:FGL002:   2.ATOM     0.000          1.699          0.000

My converge criterion is:
PORT 0.5
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7
0.6 0.6 0.6 0.7


The case.outputM have a  " ***** FALSE CONVERGENCE ***** " message.

I wonder how can I do in these cases. How can I improve the forces converge?

Thank you for reading.

Best wishes,

Jian