[lammps-users] nve or nvt for rigid bodies

Dear lammps users

I have a system of two rigid large molecules in a solution of several 10000 atoms. The large molcules are fixed at certain positions but allowed to rotate. The solution is kept at 300 K using fix/nvt for the large bodies I use fix/rigid. Since the masses differ by a factor of 1000 (between atoms and rigid particles) it is cumbersome to thermalize the large molecules. Using fix/rigid (nve) the macromolecules temperature shows large oszillations with peaks up to 1500K which are not reduced after 500000 integration steps. However, when I use a temperature scale to force the large particles temperature to oszillate around 300 K, I probably disturb my system to much. I want to measure the most preferred orientation of two macromolecules with respect to each other for fixed distances at room temperature. So my question is what kind of integrator nve or temp. scaled nve should I use to obtain the physical most reasonable results. Has anybody an idea or experiences in simulations of large rigid bodies in solution?

Thank you.



What do you mean by the "temperature" of 1 or 2 rigid bodies.
They each only have 3 degrees of freedom. What are you
expecting them to do other than wiggle their orientation?
If you're expecting LAMMPS to compute their "temperature"
how are you calculating it?