[gmx-users] Re: constraints problem

[Mark Abraham]

wurl wrote:
> thanks David,
> 	i mean i always met with problems simulating amorphous and crystalline systems without bond constraints, your suggestion of reducing the time step is very reasonable, and i don't know a compressibility of 1e-6 will be suitable for a polyethylene crystal cell?

Consult literature here.

> 	and i have also not succeeded in simulating such systems with constraints= all-angles, it always makes long warnings during the md process and stop soon. my time step is 1fs, the oplsaa force field, and the version is gromacs3.3 with pme.c updated. do i need to make specific changes to the default run perameters for systems with all angles bonds constraint?

LINCS (the default constraint algorithm) is not suitable for structures
that require "coupled constraints", including cases of constrained
angles between constrained bonds and possibly constrained angles with
common bonds. Read the manual for details. Try SHAKE.

Mark