AW: RATTLE algorithm

From: Norman Geist (
Date: Tue Feb 26 2013 - 00:38:53 CST

Hi Mihaela,

the RATTLE algorithm belongs to the "rigidbonds", needed to remove some
degrees of freedom from the hydrogens, to prevent them from oscillating and
so save computation time, as nobody cares this osccilation (btw would nobody
save the frames so fine grained to still see it later). The error you get,
points out, that a hydrogen moved that far within one timestep, that it
couldn't be restrained to its original distance with its mother atom in the
maximum iterations or tolerance. And this, finally, happens when hydrogens
move to fast. This can be caused by either a bad initial structure (a longer
minimization should help than) or a too large timestep ( > 1), maybe also
due some other things, but you should check your minimization output, if
there were some messages about bad contacts. Also you can plot your
minimization output file with VMD (NAMDplot), to check if your TOTAL energy
converged (export it to xmgrace, and zoom in), if not, try more minimization

Good luck

Norman Geist.

> -----Ursprüngliche Nachricht-----
> Von: [] Im
> Auftrag von Mihaela Drenscko
> Gesendet: Dienstag, 26. Februar 2013 00:56
> An:
> Betreff: namd-l: RATTLE algorithm
> Hi,
> I am running a simulation and I get the following error:
> ERROR: Constraint failure in RATTLE algorithm for atom 82!
> ERROR: Constraint failure; simulation has become unstable.
> ERROR: Exiting prematurely; see error messages above.
> How do I fix it?
> Thanks,
> Mihaela

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:22:59 CST