Dear Steve

thanks a lot for your quick response,

with a complex workaround in the input file we have been able to exclude the intra-molecular lennard-jones so that we can avoid the card “neigh_modify exclude”, as you said everything is correct in this case and cutoff independent.

if anyhow we do not exclude lj, we sum up so many huge positive numbers that we end-up in a numerical error bigger than the

quantities that we need to compute.

Thus it is fundamental for us to find a way to

-keep the intramolecular coulomb (we can the easily subtract it afterwards as you said and it is small)

-BUT we have to avoid the intramolecular LJ

is there an easy way to do this?

Thanks in advance for any further suggestions

Andrea

up to now what we did is:

double each molecule, defining two additional types

and excluding interactions among the types that are lj in the same molecule:

units real

atom_style full

boundary p p p

read_data doppie.dat

mass 1 12.011

mass 2 0.0001

mass 3 0.0001

mass 4 0.0001

group lelj type 3 4

velocity all create 50.0 87287

pair_style lj/cut/coul/long 30.0 14.0

pair_coeff 1 1 0.0 1.0

pair_coeff 1 2 0.0 1.0

pair_coeff 1 3 0.0 1.0

pair_coeff 1 4 0.0 1.0

pair_coeff 2 2 0.0 1.0

pair_coeff 2 3 0.0 1.0

pair_coeff 2 4 0.0 1.0

pair_coeff 3 3 0.0298 3.4

pair_coeff 3 4 0.0298 3.5

pair_coeff 4 4 0.0298 3.6

kspace_style ewald 1.0e-10

neighbor 0.3 bin

neigh_modify every 20 delay 0 check no one 100000 page 1000000

neigh_modify exclude molecule lelj

fix 1 all rigid molecule langevin 50.0 50.0 1000.0 1948

dump 1 all xyz 500 pos_50_28.04.xyz

thermo 100

thermo_style custom step temp ke pe etotal

thermo_modify format 5 %22.14g

run 100000