pair style hybrid with table

LAMMPS LAMMPS Mailing List Mirror
#1

Does anyone see a reason why I am getting this error: Coulomb cutoffs of pair hybrid sub-styles do not match.I set the cutoff to be 15 for coul/long. I did not set Coulomb interactions anywhere else.

Then if I remove coul/long and just add ewald to so linear pair style, simulation will run. Though my results seem off, I Would like to understand why I get the error when doing the 3 pair style together.

units metal
atom_style full
kspace_style ewald 1.0e-6
#Data format is read in below
read_data …/…/quartz.data

pair_style hybrid/overlay morse 2.5 coul/long 15 table linear 701
pair_coeff 1 1 morse 0.042395 1.379316 3.618701 2.5
pair_coeff 1 2 morse 0.340554 2.006700 2.100000 2.5
pair_coeff 2 2 morse 0 .1 0
pair_coeff * * coul/long
pair_coeff 1 1 table table.repulsion Repulsion_1_1
pair_coeff 1 2 table table.repulsion Repulsion_1_2
pair_coeff 2 2 table table.repulsion Repulsion_2_2

#2

Currently, LAMMPS has no way of knowing whether your tabulated
potentials encode Coulomb interactions or not. Since using
a long-space solver with pair hybrid requires that the short-range
cutoffs for all Coulomb-related potentials be the same, it includes

the pair table cutoffs in that test. Thus if your tables are not
cutoff at 15 Angs, you will get the error. Setup your tables
with that cutoff at 15 (even if that means interactions from 5 to 15 Angs have

zero force), and it should work.

I suppose a better solution is to have some flag on the individual
table potentials to indicate they are Colomb related or not.
But no one has coded that yet.

Steve

#3

Ok thanks works like a charm