[DFTB-Plus-User] Failure of routine zhegvd

[hfox1 at nd.edu]

Dear DFTB+ users,

I have just started using DFTB+, and am getting the following error a lot:

ERROR!
-> Failure in diagonalisation routine zhegvd, non-positive definite overlap!
Minor  ...  responsible.

I understand that this can be caused by atoms too close to each other. But in
one simulation I compressed my unit cell by only a small amount from
equilibrium and received the error. Is this normal or is there sometimes
another reason behind the error?
I also have gotten the error when I do nothing except change the kpoints.
At the moment I'm simulating a unit cell of titanium dioxide.

Thanks,
Hannah Fox