AMBER Archive (2002)

Subject: Re: Density during equilibration (sander)

From: Marc Dobler (marc_at_mummy.tokai.jaeri.go.jp)
Date: Tue Sep 17 2002 - 19:29:22 CDT

> > I can try it, I was avoiding longer NPT simulations because the goal in
> > the longer run is to study protein dynamics. In a previous post you had
> > mentioned that temperature baths and pressure pistons distort the
> > system and thats why I was avoiding longer NPT simulations.
>
> You should make a distinction between equilibration and production runs.
You
> need enough NPT equilibration to get a good density and temperature, but
you
> don't analyze results from that. After getting the right density, you can
> switch to NVE for the production part, which is where you look at
dynamics.
>

i also encountered this problem.
first, i make a long equilibration run with NTP=1 (and NTB=2),
for which the pressure is 1 and the density reaches efficiently quite well
1.
but when starting the production part,
if i switch to constant volume, the density decreases again to around 0,94.
so, i need to keep constant pressure during the production part.

i have not found yet any other way to keep pressure and density stable.
maybe somebody has any good idea to solve this.

Marc Dobler