[gmx-users] Tabulated potentials

[Laura Leay]

All, 

I'm simulating a very small system of methane using tabulated
potentials. Charges are not modified and I'm using a cubic for values up
to 0.19 nm and the normal 6-12 LJ potential after that (the form the
cubic is such that the value and gradient at 0.19 nm is the same as for
the 6-12 LJ). 

The potential is different for each pair interaction so I have specified
separate tables for the C-C and H-H interaction leaving the default
table.xvg to deal with the H-C interactions. So in my mdp file I've
specified:
  coulombtype = Cut-off
  vdw-type = user
  energygrps = CT HC ;(consistent with the conf.gro file)
  energygrp_table = HC HC CT CT

My table files are:
- table.xvg
- table_CT_CT.xvg
- table_HC_HC.xvg

I have specified an index file for the system, the hydrogens and the
carbons in the file index.ndx.

Each table file gives values down to 3.0 nm with 0.002 nm spacing. I am
using gromacs version 4.0.7 single precision.


However when I try and grompp and run the simulation using:
  grompp -v -n index.ndx
  mdrun -v table table.xvg

I get the following error:

  Tables in file table.xvg not long enough for cut-off:
	should be at least 1.900000 nm


I've compared my table files to the examples given in gromacs's own
topology directory and the increments between each data point is exactly
the same (0.002 nm). I've also checked that the increments for each of
my table files are consistently 0.002 nm. I'm not sure what else to
check to resolve this error.

Help please.