Re: "Constrain failure" in NAMD

From: Edward Patrick Obrien (
Date: Fri May 07 2004 - 17:59:05 CDT

Hi Mike,
  First of all make sure you do these things first:
        [0] CHECK THAT YOUR DIHEDRALS/BOND ANGLES are defined in your .psf
        file! ( I got this error "ERROR: Constraint failure in RATTLE
        algorithm for atom " before and it was because of a screwed up
        [1] Minimize until 'gradient tolerance' is below 1 or so.
        [2] Heat the system from O K to your temperature of interest
        slowly. (e.g. I heat mine from 0K to 300 K over 300 000 ps )

  Check and do these things before anything else is my suggestion.


PS- Rattle is used to correct the position of atoms attached with rigid
bonds. When integrating newton's equations of motion with the verlet
algorithm over atoms that have rigid bonds a certain error in the postions
is introduced... rattle tries to correct this error. I think thats about
right, you can look it on the NAMD site too.

On Fri, 7 May 2004, Michael Chen wrote:

> Hello, Everyone,
> Hope all is well for you.
> It was so strange for me that when I used "run" command, I got lots of
> ERROR messages as following:
> ERROR: Constraint failure in RATTLE algorithm for atom 2264!
> ERROR: Constraint failure; simulation has become unstable.
> What did this mean? How should I kill this problem? I minimized my
> system for quite a long time, and I turned the constraint on for the whole
> backbone.
> Thanks for the answer.
> May you have a nice day.
> *****************************************************
> Michael Chen
> Department of Biochemistry
> SUNY Upstate Medical University
> 750 East Adams Street
> Syracuse, NY 13210
> *****************************************************

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:37:35 CST