[gmx-users] strange temperature coupling problem

[Nickle Fan]

What I meant by using a tau_t of 3000 ps is to effectively turn off the
thermostat for water. What I want to do is let the wall thermostat to take
care of the temperature control (if the water temperature rises, the heat
are supposed to get out through the wall thermostat). But somehow this does
not happen.

Nick

On 1/12/06, Simple <qsb at nlbmol.ibp.ac.cn> wrote:
>
> Nickle Fan writed£º
> > Dear Gromacs users:
> >
> > I am having a rather strange problem in one of my simulation. The system
> > consists of 2000 water molecules embedded between two walls. The wall
> atoms
> > are position restrainted (force constant: 5000).
> >
> > The system is first equiliburated under NVT (Berendsen thermostat, tau_t
> =
> > 0.1ps, Tsystem = 300K) for a while. Then I coupled the wall atoms to
> 300K
> > (tau_t = 0.1 ps), and water to 300K (tau_t = 3000 ps). I found that
> during
> I found you used (tau_t = 3000 ps) for water, I don't know why you use
> different tau_t for wall and water.But I think, under your setting, the
> temperature will only couple at  3000 ps interval. So in a 100ps
> simulation,
> the temperature of water would never adjust,Then the increasing of temp
> seem
> reasonable.
> I'm not familiar with this problem.I only notice the strange setting,
> maybe
> there are other opinions,
> > a 100ps simulation, the temperature of water rise continuously to 400K!
> > (the wall atoms are OK though). I tried to disable water temperature
> > coupling by set tau_t = 0 for water, but it does not help.
> >
> > I tried NVE simulation from the same initial configuration, and it
> appears
> > that the temperature of water during the first 50 ps is about 300K. So
> we
> > can rule out the possibility of bad local contact in the system. Then
> what
> > could possibly goes wrong with the system?
> >
> > I  am using PME (4th order), vdw-type = Shift, rvdw-switch  = 0, rvdw =
> > 1.1, rlist = 1.2. Time step is 1fs.
> >
> > Any help would be greatly appreciated.
> >
> > Nick Fan
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20060112/2615e94a/attachment.html>