From: Vlad Cojocaru (Vlad.Cojocaru_at_mpi-bpc.mpg.de)
Date: Fri Jan 23 2004 - 02:54:03 CST
I know that for instance in Amber a good value for the constant is
about 2-5ps for production runs (the coupling is weaker) while for
equilibration if you want tighter coupling one can use 0.5ps. So for
NAMD I use 1/ these values......but I am not completely sure about the
equivalence of tautp (amber) and 1/const (NAMD).
(you could also try to apply trial/error by using a range of constants
from 0.150 to 1.000 and then you will see what's the effect....the
bigger the constant the faster you will reach the desired temp.
Hope this helps,
Jim Phillips wrote:
>The coupling constants have units of 1/ps, so you would put 2.5 in the B
>column of the PDB file. Here's a snip of the relevant source code:
>void Sequencer::tcoupleVelocities(BigReal dt_fs, int step)
> if ( simParams->tCoupleOn )
> FullAtom *a = patch->atom.begin();
> int numAtoms = patch->numAtoms;
> BigReal coefficient = broadcast->tcoupleCoefficient.get(step);
> Molecule *molecule = Node::Object()->molecule;
> BigReal dt = dt_fs * 0.001; // convert to ps
> coefficient *= dt;
> for ( int i = 0; i < numAtoms; ++i )
> int aid = a[i].id;
> BigReal f1 = exp( coefficient * molecule->langevin_param(aid) );
> a[i].velocity *= f1;
>This is a pretty old method. I'd suggest using Langevin personally.
>On Tue, 20 Jan 2004, Hyonseok Hwang wrote:
>>I would like to use the 'Berendsen's temperature coupling method' in
>>NAMD. In that method, a coupling constant of 0.4ps is generally
>>recommended. Then my question is what value I should put in the B
>>column, which is now designated as the 'tCouplecol'. Do I should put 0.4?
>>Thank you very much.
-- Vlad Cojocaru Max Planck Institute for Biophysical Chemistry Department: 060 Am Fassberg 11, 37077 Goettingen, Germany tel: ++49-551-201.1327 e-mail: Vlad.Cojocaru_at_mpi-bpc.mpg.de home tel: ++49-551-9963204
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:38:22 CST