Once again about periodic colvar

[ilya]

Hello!

Well, I performed another test reaction which is Cl- + CH3Cl (from the
article).
I have successfully build a 3d fes which is very similar to that one
in the article and I think I've got the idea of choosing parameters
for metadynamics.
   BUT
For the case of butane I can't obtain any adequate free energy
profile.
I observed that after some point  the molecule starts spinning
strangely around C2-C3 bond and I think the are still some bugs in the
case of periodic colvar in code.
Maybe someone else with more experience in metadynamics would also try
to do this? Maybe I am doing something wrong?


On May 14, 4:42 pm, Teodoro Laino <teodor... at gmail.com> wrote:
> Ciao Ilya,
>
> Thanks again.. hopefully we should soon converge to a bug-free  
> metadynamics for periodic colvar ;-)
> There were few other missing adjustments, now in CVS.
> At least now also for extended lagrangian the ss0 and ss should be  
> both (for periodic colvar) in the
> range (-pi,pi].
>
> Check it and let us know.
> Thanks,
> Teo
>
> p.s.: of course the role of the parameters for metadynamics is very  
> important. And your attitude is the
> correct one. Keep trying!
>
> On 14 May 2008, at 09:16, ilya wrote:
>
>
>
> >  Hello!
>
> > When working with periodic colvar I noticed that while the real value
> > of colvar (instantaneous colvar value) is always printed within the
> > range (-pi,pi], but the colvar value of the extended lagrangian is not
> > corrected for the period. So for example
>
> > lagrange cv      inst val
> > 5.23293          -1.08931
> > 5.23505          -1.08060
> > or
> > 13.27589         0.70540
> > 13.27910         0.70803
>
> > Should the lagrange cv be corrected for N*2*pi?
> > The same spawned hills positions are printed to restart file.
>
> > Looking at the code of the fes program I didn't found the place where
> > the lagrange vals are corrected.
>
> > The main reason I posted this is that I still can't get the correct
> > free energy profile for butane rotation. I've played a lot with
> > parameters of metadynamics simulation (according to the article "A
> > Recipe for the Computation of the Free Energy Barrier and the Lowest
> > Free Energy Path of Concerted Reactions" and some others).
>
> > Could the above thing be the source of errors or I should continue
> > playing with params.
>
> > Thanks.
>
> > Ilya.