Re: Constraint Failure in SMD simulation

From: Ramin Omidvar (
Date: Tue Jul 13 2010 - 08:04:20 CDT

Dear felmerino

I tried to equilibrate my system by applying more run step before running
SMD simulation, I minimized and equilibrated 3 ns totally but when I run
SMD, error was as same as before, "constraint failure ...."
And after decreasing time step(Tricking with TS) , the simulation started to
run at a TS that was similar to my earlier simulations. but It failed too.
I also decreased applied force but same TS allows simulation to run.
I would like to know that this error may related to my protein and system.
Is there anything that cause this error or still time step for equilibration
is not enough, Any hint and guide is highly appreciated,

Best regards
Ramin Omidvar
PhD student of Biomedical Engineering Dept.
AmirKabir University of Tech.

On Thu, May 20, 2010 at 12:18 AM,

> Ramin,
> That error means that indeed your system is crashing due to the
> perturbation you are applying. As you noticed short time steps do the trick,
> Maybe the energy function for your system is too sharp and you'll need to
> pay that price. Anyways, decreasing the applied force also could solve your
> problem. Of course supposing that you system is well equilibrated, otherwise
> it will always crash.
> ----Mensaje original----
> De:
> Fecha: 19-05-2010 12:51
> Para: <>
> Asunto: namd-l: Constraint Failure in SMD simulation
> Dear NAMD users,
> Hi all,
> I am running a SMD simulation to study the unfolding behavior of CNTN
> protein based on tutorial's configuration file in "constant force pulling"
> section but I have been faced with below error for 7 atoms (All atoms are
> belong to CNTN molecule not to the solvent water molecules).
> ERROR: Constraint failure in RATTLE algorithm for atom 4719!
> ERROR: Constraint failure; simulation has become unstable.
> ERROR: same as above for other six atoms
> I played with TimeStep and decreased it several times and finally the
> error was removed but I was forced to increase the number of steps and
> total run time was not as enough as other usual simulations so I think that
> it is not the most appropriate solution.
> I would be so appreciated if somebody could show another way to me
> or confirm my solution.
> Any assistance is highly appreciated ,Thanks in advance,
> Best regards
> Ramin Omidvar
> PhD student of Biomedical Engineering Dept.
> AmirKabir University of Tech.

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:54:18 CST