Re: [Towhee-users] Help with setting rigid dihedral torsion interaction in a 12-site rigid model of benzene

[Marcus M. <mar...@us...>]

On Wed, Apr 8, 2009 at 10:32 AM, Anatoli Milischuk <to...@gm...> wrote:
> I'm trying to set up a rigid 12-site benzene for grand canonical or NPT
> ensemble simulation.
> I am building the initial configuration using the template file. The bond
> lengths and angles are fixed.
> I use multiple rigid dihedrals (Torsion styel 21) to set up rigid torsion
> angles.
> When I run the simulation with the attached files I get this output error
> (see output file for detail)
> " Initial Energies for Box     1
>  overlap torsion
>  imolty,ttype            1            7            1            2
> 3
>  MAINLOOP: overlap in initial configuration "

The first problem is you are only allowing a dihedral value of 0.0.
This is fine for benzene in the cases of H-C-C-H and C-C-C-C, but it
does not work for the H-C-C-C torsion as that should have an allowed
rigid value of Pi (about 3.1415 in your benzene_ff file).

> In the force field file benzene2_ff for Torsion Type Number 2
> I set up the torsion parameters to 0.001 (deviation from the rigid angle)
> and 0. (dihedral angle).
>
>
> When I change the deviation from the rigid angle parameter to something >
> 4.d0 I am able to run my simulation but then I'm getting a non-zero value
> for
> Intra vdw energy at the end of the run.

Increasing the deviation that large means that almost all dihedral
angles are allowed by a multiple rigid torsion potential that has an
allowed deviation of 4 as the units are in radians.

There was also a lot of overlap as you were putting the 10 molecules
onto an 8 by 8 by 8 grid initially.  If you change inix iniy iniz to 3
3 3 that will help you out.

Marcus
--
Marcus G. Martin
Director, Useful Bias Incorporated
88 Martinez Road
Edgewood NM 87015-8222
ph. (505) 286-4457
www.usefulbias.com