[gmx-users] Planar molecules / virtual sites

[Dimitris Dellis]

Hi All,
I try to simulate with gromacs substituted ethlyle series of molecules 
(planar) and I have to keep molecules rigid.
In case of atomic substitutes, I use 3 virtual sites on the plane of 
molecule, that reproduce mass and moment of inertia,
I use 3 constrains to keep them rigid and the massless molecule atoms 
are expressed as 3fd sites.
Sometimes, depending on masses and bond lengths parameter  a of 3fd is > 1.
Shake fails.

Any suggestion how to handle this type of molecules ?