[lammps-users] fix rigid/nvt for CO2

Regarding unit conversion, the only difference I can see from the code so far is the force->ftm2v variable, which comes in the timestep used to update conjugate quaternion momentum, dtf:

for 'metal' units: force->ftm2v = 1.0 / 1.0364269e-4 = 9648.533823;
for 'real' units: force->ftm2v = 1.0 / 48.88821291 / 48.88821291 = 4.1839999e-4;

(see Update::set_units() in update.cpp)

I guess this difference might affect the integration stability even when you are using the same physical timestep.


Quoting Steve Plimpton <[email protected]>: