[gmx-users] water molecule not settled, is due to the change in the form of the box?

[Attilio]

Dear all,

I'm doing NPT md on my system, a small drug non covalently linked to DNA, and I already performed 8.5 ns simulation.
At this point I get this message 

"Inertia tensor (3x3):
   Inertia tensor[    0]={ 4.61676e+03,  1.35627e+03, -2.24899e+03}
   Inertia tensor[    1]={ 1.35627e+03,  8.25089e+03, -3.82382e+02}
   Inertia tensor[    2]={-2.24899e+03, -3.82382e+02,  1.18415e+04}

t = 8898.120 ps: Water molecule starting at atom 11315 can not be settled.
Check for bad contacts and/or reduce the timestep.Wrote pdb files with previous and current coordinates

Step 5932081, time 8898.12 (ps)  LINCS WARNING
relative constraint deviation after LINCS:
max 0.000000 (between atoms 1 and 2) rms nan
bonds that rotated more than 30 degrees:
 atom 1 atom 2  angle  previous, current, constraint length

t = 8898.121 ps: Water molecule starting at atom 11993 can not be settled.
Check for bad contacts and/or reduce the timestep.Wrote pdb files with previous and current coordinates

Step 5932082, time 8898.12 (ps)  LINCS WARNING
relative constraint deviation after LINCS:
max 0.000000 (between atoms 1 and 2) rms nan
bonds that rotated more than 30 degrees:
 atom 1 atom 2  angle  previous, current, constraint length

t = 8898.123 ps: Water molecule starting at atom 11993 can not be settled.
Check for bad contacts and/or reduce the timestep.Wrote pdb files with previous and current coordinates
 ....
.....
....

t = 8898.141 ps: Water molecule starting at atom 11993 can not be settled.
Check for bad contacts and/or reduce the timestep.Wrote pdb files with previous and current coordinates
Grid: -2147483648 x -2147483648 x -2147483648 cells
Fatal error: ci = -1 should be in 0 .. -1 [FILE nsgrid.c, LINE 218]"


I checked the simulation, and I see that the box, originally cubic, transform to a rectangular box, so the minimum distance
between the images decrease abruptly from 35 to 15 angtroms, and the DNA strands separate (it's beatiful to see, 
not to report to your boss...).

I'm using a time step of 1.5 ps, and a cutoff of 12 angstrom for both electrostatic and van der waals.


How may I go on without break my system?

Thanks in advance, 
Attilio