[DFTB-Plus-User] Failure in diagonalization routine

[Jacek Jakowski]

You should print the overlap matrix to the file. Then read in the overlap
matrix  to   matlab,  octave or something like.   Then you should
diagonalize  your  overlap. You should see  all the eigenvalues to be
positive and close to 1. If you see  something  very close to zero  the you
have a linear dependence in your basis set.  If you see something
negative,   then   you clerarly have your reason  for cholelsky failing.
You  can  then inspect  the corresponding eigenvectors.

Jacek

On Thu, Mar 26, 2015 at 6:35 AM, Yusupov Maksudbek <
maksudbek.yusupov at uantwerpen.be> wrote:

>  Dear Dr. Balint,
>
> I am a new user of DFTB. My system consists of biomolecule, such as
> peptide and I want to study the interaction of OH radicals with this
> biomolecule.
> The peptide is located at the center of the periodic box. The box is big
> enough, i.e., minimum image criterion was taken into account. Ten OH
> radicals are randomly positioned around the structure (with minimum 5 A
> distance from the structure and from each other).
> When I run the simulation I have this error “Failure in diagonalization
> routine dsygvd, non-positive definite overlap! Minor 107 responsible” after
> several steps.
> I have checked this orbital (i.e, 107) and the atom is not unphysically
> close to other atoms.
> Please find the dftb_in.hsd and detailed.out files in the attachment and
> please help me to solve this problem.
> I appreciate your time and response in advance.
>
> Kind regards,
> Maksud.
>
>
>  **************************************************
> dr. Maksudbek Yusupov
> Department of Chemistry, Research group PLASMANT
> University of Antwerp
> Universiteitsplein 1
> BE-2610 Wilrijk-Antwerp, Belgium
> Tel: +32 3 265 23 82
> Fax: +32 3 265 23 43
> E-mail: *maksudbek.yusupov at uantwerpen.be
> <maksudbek.yusupov at uantwerpen.be>*
> Website: *https://www.uantwerp.be/en/rg/plasmant/
> <https://www.uantwerp.be/en/rg/plasmant/>*
> **************************************************
>
>
> _______________________________________________
> DFTB-Plus-User mailing list
> DFTB-Plus-User at mailman.zfn.uni-bremen.de
> https://mailman.zfn.uni-bremen.de/cgi-bin/mailman/listinfo/dftb-plus-user
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20150326/3a2229a2/attachment.html>