[gmx-users] (no subject)

[Erik Lindahl]

Atom News wrote:

>
> Dear GROMACSER,
>
> I have simulated the protein in water. After 362ps, I got the message:
>
> Water molecule starting at atom 1094 can not be settled.
> Check for bad contacts and/or reduce the timestep.
>
> Can you help me? Thanks. 


This happens when an atom in a water molecule moves too fast (i.e. the 
water rotates too much in a timestep.), so the SETTLE algorithm that 
constrains the water bond lengths and angle breaks down. It CAN happen 
just by chance once in a blue moon in very large systems (in that case 
it will probably help if you restart from the last position in your 
trajectory, simply because you'll take a *slightly* different path in 
phase space the next time). Usually it's because your timestep is too 
large though (more than 2-3 ps), or because the water molecule is 
bumping into something.

Cheers,

Erik