[lammps-users] Long FENE bonds

Hi: Bu default, when a FENE bond length becomes larger than 90% of its
upper limit, LAMMPS will issue an warning. I modified the file
bond_fene.cpp and changed 90% to 80%. I set R_0 1.5 and was hoping
that LAMMPS would issue warnings when any bond gets longer than 1.2.

For one test run, in the output file, there is no warning at all. But
when I calculated the bond lengths by myself using the atom position
and connectivity information, I found a few bonds with lengths larger
than 1.2. Is it possible that LAMMPS fails to detect and issue
warnings on long FENE bonds ?



I also have some issues related to FENE bonds. If repulsive LJ polymer
chains are used, then it works fine, after running Soft potential for
a million steps. If I use attractive LJ chains then it creates 'long
FENE bond' problem. Steve said earlier, system needs to be properly
equilibrated. So, does the system needs to be equilibrated with
repulsive Soft + repulsive LJ, and then do the main run with
attractive LJ, or just Soft potential is enough before the main run.
As I understand Soft potential basically pushes the overlapping
monomers away.


Hi, Ting. It’s possible but unlikely - I did this change to 80% earlier with no problems.


Warning messages in the middle of a run go to
the screen but not the log file. This is to avoid
polluting the thermodynamic output in the log file,
making it hard to post-process. If you did the
change correctly, this is something LAMMPS checks
every timestep (when bond forces are computed),
so you should see it.


Soft is enough by itself (in systems I've worked with), if
you do it long enough and if the "soft" prefactor becomes hard enough
to push things away.


Hi: Thank you. I found the reason why my calculation on bond lengths
was different from the LAMMPS. I did not pay attention to the fact
that by default the dump atom command dumps the atom coordinates in a
scaled format. And I forgot to rescale the coordinates.
After rescaling, the results are consistent.