Re: [ESPResSo-users] VIRTUAL_SITES

[Rudolf Weeber]

Hi,
On Thu, Aug 21, 2014 at 04:50:33PM +0200, Dudo wrote:
> please, I would need a sort of under the hood information,
> do the virtual sites have mass and gamma or only if VIRTUAL_THERMOSTAT is
> called?
The mass property of virtual sites is ignored, as they are not propagated using 
Newton's eqn of motion.
Rather, their position is calculated from the position and orientation of the 
particle, they are related to, i.e., the particle in the center of mass of the 
rigid body.
Unless you set the switch mentioned above, virtual sites are not thermostated. 
Instead, the thermostating is done through the rotational and translational 
degree of freedom of the particle in the center of mass of the rigid body.
If you want correct dynamics, you will probably need to set
mass of the particle in the rigid body's center of mass =total mass of all 
particles forming the rigid body.
Interial tensor fo the particle in the center of mass = Inertial tensor of the 
rigid body as a whole.
Here, it is important to make sure that the orthogonal basis in which you 
specify the inertial tensor co-insides with the co-rotating frame of reference 
of the particle in the center of mass.
The easiest way to achieve this, is to set the quaternions of the central 
particle to 1 0 0 0.
In this case, the co-rotating frame of reference is equal to the lab frame.

If you only need static properties, you can ignore the mass and  inertia of the 
particle in the center of mass.

It is currently not possible to set the gamma parameter of the Langevin 
thermostat independently for the rotational and translational degree of freedom.

Hope that helps.
Regards, Rudolf