Re: Can a small langevinDamping coefficient lead to 'ERROR: atoms moving too fast'?

From: Jose Borreguero (
Date: Tue Aug 27 2013 - 16:25:38 CDT

I equilibrated the system for 2ns and then ran in the NVE. As it turns out,
energy is not conserved but increases rapidly! I don't know what I'm doing
wrong, I have set 1fs time steps, no rigid bonds, update the electrostatics
and pair list every fs!

One possible cause might be the non-cubic periodic boundary conditions. It
could be that I have clashes between atoms in adjacent cell images because
the boundary conditions are incorrect. How could I check for this

Here's my input file for NVE simulations:

amber on ;# use Amber FF, expect to read a PARM7
switching off ;# do not use switching
exclude scaled1-4 ;# 1-4 interactions are scaled by scnb and
1-4scaling, 1-3 interactions are ignored
1-4scaling 0.833333333 ;# scaling factor for 1-4 electrostatic
scnb 2.0 ;# 1-4 vdW interactions are divided by
readexclusions yes ;# read exclusion lists from the PARM7 file
LJcorrection off ;# analytical tail correction to reported
vdW energy
ZeroMomentum on ;# remove center of mass drift due to PME

set inputname eq0 ;# prefix for input files
set outputname eq1 ;# prefix for output files

parmfile crystal.prmtop ;# topology (AMBER)
coordinates crystal.pdb
bincoordinates $inputname.restart.coor
binvelocities $inputname.restart.vel # set this if no temperature is

margin 0.5 ;#(default 0.0) extra length in patch dimension, neccessary when
box volume changes fast

cutoff 12.0 ;# may use smaller, maybe 10., with PME
switchdist 10.0 ;# cutoff - 2.
pairlistdist 14.0 ;# cutoff + 2. Promise that atom won't move
more than 2A in a cycle
stepspercycle 1 ;# redo pairlists every X steps (10 for
equilibrium simulations)

timestep 1 ;# 1fs/step
rigidBonds none ;# needed to record vibrations of bonds
involving hydrogens
nonbondedFreq 1 ;# nonbonded forces every step
fullElectFrequency 1 ;# PME only every other step

extendedSystem $inputname.restart.xsc ;# extended info file

wrapWater off ;# wrap water to central cell
wrapAll off ;# wrap other molecules too
wrapNearest off ;# set to 'on' for non-rectangular cells when using
wrapWater or wrapAll

PME off
PMEGridSpacing 1.0
PMEGridSizeX 45
PMEGridSizeY 45
PMEGridSizeZ 45

outputName $outputname ;# prefix for output files
restartfreq 10000 ;# overwrite restart file every Xsteps
DCDUnitCell yes ;# write unit cell data to DCD file
dcdfreq 1000 ;# write coords every Xsteps
xstFreq 80000 ;# write extended trajectory every Xsteps
#forceDCDfreq 100 ;# trajectory of atom forces every Xsteps
outputEnergies 10000
outputPressure 10000

firsttimestep 0 ;# reset frame counter
run 8000000 ;# 8ns


On Mon, Aug 26, 2013 at 12:56 PM, Gianluca Interlandi <> wrote:

> A few of the simulations aborted with the following message:
>> "ERROR: Atoms moving too fast; simulation has become unstable."
>> Could a small damping coefficient lead to such errors? If so, what would
>> be the scenario
>> giving rise to some atom moving too fast?
> Atoms move too fast if there are huge forces which is often the result of
> an incorrect system setup, like atoms clashing or too close too each other.
> You need to perform sufficient minimization and equilibration before
> starting production. You might want to inspect your initial PDB for
> clashing atoms. Sometimes hydrogen atoms clash after they have been built.
> In this case, perform minimization and equilibration with harmonic
> constraints on heavy atoms in order to minimize and equilibrate the
> hydrogen atoms.
> Gianluca
> ------------------------------**-----------------------
> Gianluca Interlandi, PhD
> http://artemide.bioeng.**>
> Research Scientist at the Department of Bioengineering
> at the University of Washington, Seattle WA U.S.A.
> ------------------------------**-----------------------

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:34 CST