[lammps-users] coloumb energy affected by bonds

hi,
I am testing out a sodium chloride crystal with 8 atoms in a periodic box. When I run the simulation with just coloumb effects using a pppm long-range solver I get energies which are in agreement with the literature(eg. Kittel). However, I wanted to see if adding bond coeffs would affect anything. So I added one bond between an Na and a Cl with a large energy coefficient and the distance coeffient I set to equal their equilibrium distance of 2.82 Angstroms. The bond energy came out as zero which I expected but the coloumb energy changed. Am I missing something?

-Dan

My first run with no bonds:

PPPM initialization ...
  G vector = 0.459197
  grid = 32 32 32
  RMS precision = 5.38115e-09
  brick FFT buffer size/proc = 117649 32768 64827
Setting up run ...
Memory usage per processor = 9.41173 Mbytes
Step Temp PotEng E_mol E_long E_coul E_bond TotEng
       0 0 -823.12432 0 -668.87741 -154.24691 0 -823.12432
     100 1.8676036e-28 -823.12432 0 -668.87741 -154.24691 0 -823.12432
     200 8.5587672e-28 -823.12432 0 -668.87741 -154.24691 0 -823.12432

The second run with one bond added:

PPPM initialization ...
  G vector = 0.459197
  grid = 32 32 32
  RMS precision = 5.38115e-09
  brick FFT buffer size/proc = 117649 32768 64827
Setting up run ...
Memory usage per processor = 9.48809 Mbytes
Step Temp PotEng E_mol E_long E_coul E_bond TotEng
       0 0 226.17823 0 -668.87741 895.05564 0 226.17823
     100 1.814077e-27 226.17823 0 -668.87741 895.05564 0 226.17823
     200 1.1599017e-25 226.17823 0 -668.87741 895.05564 0 226.17823

Look at the special_bonds command. By default when there
is a bond between 2 atoms, the pairwise interaction between
them is turned off.

Steve