AMBER Archive (2005)

Subject: AMBER: Andersen thermostat: target T?

From: Arantxa Sanz (aranzazu.sanz_at_uab.es)
Date: Fri Jan 21 2005 - 13:06:01 CST

Hello:
 
I have been running the following Sander script to run MD dynamics at
constant T with Andersen thermostat
(ntt=2), scaling T from 0 K to 300 K in 15ps :
    MD of receptor and solvent

    &cntrl

    imin=0, nmropt=1,ntx=1, irest=0,

    ntxo=1, ntpr=250, ntave=10000,

    ntwx=1000, ntwv=0, ntwe=1000, ntwr=1000, iwrap=1,

    ntf=2, ntb=1, dielc=1., cut=10.,

    scnb=2., scee=1.2,

    nstlim=7500000,

    t=0., dt=0.002,

    ntt=2, temp0=300., tautp=0.2,

    ntc=2,

    &end

    &wt type='TEMP0', istep1=0, istep2=250,

    value1=2., value2=10., &end

.......................

    &wt type='TEMP0', istep1=7251, istep2=7500,

    value1=290., value2=300., &end

    &wt type='TEMP0', istep1=7501, istep2=50000,

    value1=300., value2=300., &end

    &wt type='END', &end

    &rst iat(1) = 0, &end

    Discarding the first 200ps, it seems that average T oscillates
around 310K. That is, 10K above the given target T. I had run the same
calcs

using Langevin dynamics (ntt=3) for scaling from 0 to 300K and then
using Andersen for the equilibration-production run, and I also have
observed

this effect. What have I been doing wrong? I have been told not to run
my production run at constant P because the solvent is represented by a
methane box, and

allegedly, constant P would not work right...

 

 
 
Arantxa Sanz
Postdoctoral Researcher
Laboratori de Medicina Computacional
Facultat de Medicina, Universitat Autònoma de Barcelona
08193 Bellaterra
SPAIN
 

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu