As Jason says, there doesn't seem to be anything wrong with what you've
got, you just have a lot of noise around the mean temp because of your
small system size. Implicit solvent won't affect this, so it is the same
as if you just had your 5 nt in a vacuum, there will be a lot of
temperature fluctuations.
If you are truly concerned, and don't want to go to explicit solvent as
Jason mentioned, you can maybe try playing around with the thermostat
coupling parameters, tautp and gamma_ln. I think that you can change these
to get tighter coupling, but this might not be ideal either, you'll have to
test it.
And yes, for implicit solvent, probably safe to go with at least 16
angstrom cutoff as the manual suggests, maybe 20 is even better.
~Aron
On Fri, Feb 24, 2012 at 3:25 PM, Jason Swails <jason.swails.gmail.com>wrote:
> On Fri, Feb 24, 2012 at 2:35 PM, amir abbasi <amir-abbasi.hotmail.com
> >wrote:
>
> >
> > Hi!
> > I'm trying to simulate a small nucleic acid molecule (5 nt) in implicit
> > solvent and constant temperature (300 K).
> > But i got a broad range of temperature from 250 K to 350 K.
> > Why?
>
>
> Because your system is very small. As the number of particles drops, the
> size of temperature fluctuations jump.
>
> And How i can solve this problem?
> >
>
> Why do you say this is a problem? What is your average temperature over
> the whole simulation? If you simulate in explicit solvent, your
> temperature fluctuations will drop. You can try playing with various
> thermostats (e.g. anderson, berendsen, or langevin) and temperature
> coupling constants, but you should be aware of the shortcomings/properties
> of each (for example, berendsen does not actually sample rigorously from
> the canonical ensemble).
>
>
> > This is my MD input file:
> > &cntrl
> > imin = 0,
> > ntb = 0,igb=1,
> > ntt = 3, gamma_ln = 1.0,
> > tempi = 300, temp0 = 300,
> > nstlim = 100000000, dt = 0.001,
> > ntpr = 100, ntwx = 100, ntwr = 1000
> > cut = 12.0
> >
>
> This is a _very_ small cutoff for implicit solvent calculations. It could
> be that your system is smaller than 12 Å, but I would still suggest that
> this value should be considerably larger. I don't know how much of an
> effect this will have on your temperature fluctuations.
>
> HTH,
> Jason
>
> --
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Candidate
> 352-392-4032
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Aron Broom M.Sc
PhD Student
Department of Chemistry
University of Waterloo
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Feb 24 2012 - 13:00:02 PST
