[lammps-users] ERROR: Shake clusters are connected

Hello All,

I am trying to restrain bonds involving hydrogen atoms using the following fix

fix 1 all shake 1e-6 500 0 m 1.0 #SHAKE is applied on Hydrogen atoms

and I keep getting the error, Shake clusters are connected.

The “common errors” page on lammps.sandia.gov describes it as: A single cluster specified by the fix shake command must have a single central atom with up to 3 other atoms bonded to it.

I then tried to run NVT to check if the bonds information is correct and the simulation ran just fine. What should/shouldn’t I do to restrain the hydrogen atoms ?

Thank for your time,

Pavan

The input script is:
units real
neigh_modify delay 2 every 1

atom_style full
bond_style harmonic
angle_style harmonic
dihedral_style harmonic

pair_style lj/charmm/coul/long 8 10
pair_modify mix arithmetic
kspace_style pppm 0.00008

read_data data.in

special_bonds amber
thermo 1000
thermo_style multi
timestep 1

fix 1 all shake 1e-6 500 0 m 1.0 #SHAKE is applied on Hydrogen atoms
fix 2 all npt 312.0 312.0 100.0 xyz 1.0 1.0 1000.0 #Performing a NPT simulation with an isotropic pressure of 1atm.
velocity all create 0.0 12345678 dist uniform

restart 1000 proj.restart1 proj.restart2
dump 2 all atom 1000 proj_nvt.dump
dump_modify 2 image yes scale yes
run 300000

Hi Pavan. My guess is that the two H’s on your waters have a bond between them. Take a look at the rhodo example in the “bench” directory. Here’s the shake command:

fix 1 all shake 0.0001 5 0 m 1.0 a 232

Where the “a 232” part tells LAMMPS to constrain the water angle to be rigid. In CHARMM (& AMBER?), you constrain the “bond” between the water H’s to be rigid, but in LAMMPS, you don’t do that but rather constrain the angle between them to be rigid. In the rhodo example, note that there are only 2 bonds between the last water (and all of the waters):

27722 112 31967 31968
27723 112 31967 31969

If my guess at what the problem is turns out to be incorrect, send me your data.in file and I’ll try running it here. If the data file is larger than 200 kb, you won’t want to send it to the list, but just to me.

Paul