[lammps-users] fix rigid error

Hi LAMMPS users,

When I used fix rigid/nvt rod molecule to simulate rod-coil copolymer
system by DPD, I got an error message at the last step. When I
checked the documentation. I got the following message:
Fix rigid: Bad principal moments
The principal moments of inertia computed for a rigid body are not
within the required tolerances.

I have three questions:

1. What is the value of the required tolerance?
2. Why did I get the error message at the last step? since I can run
the simulatuons untill the last step, which is to write a restart
file.
3. I am using DPD, which allows overlapping of particles. Is it normal
to get "Dangerous builds"? I checked the ouput for E_vdwl. It is about
7.

Thank you for your help.

Dongsheng

For the fix rigid problems, you should send your test files that
reproduce the problem to the author of fix rigid/nvt, who is
listed at the top of the source file. They may have an idea.

Re: dangerous builds, that has to do with neighbor list
creation. If you use neigh_modify check yes delay 0 every 1
you are as conservative as possible about re-building
lists, and should get dangerous = 0. If not, your timestep
is too big, or neigh skin is too small (by a large margin).

If you have bad dynamics (which dangerous > 0 may be
indicating), then that could also mess up rigid body motion.
So I would fix that problem first and see if the rigid issues
go away.

Steve