As mentioned in your previous discussion and written in the doc, the Tdamp is temperature
damping parameter in time units, e.g. psec for unit metal, for both 'fix nvt' and 'fix npt'.
So, if set thermostatting as:
........
unit metal
velocity all create 600 222
fix heat all npt temp 600 600 0.2 iso 0.0 0.0 1
........
then temperature will ramp to 600k in about 200 timesteps, but my simulation shows the damping
time is not about 200 but 2000 timestep. The 'timestep' command is 0.001 by default.
Here are my results:
As mentioned in your previous discussion and written in the doc, the Tdamp is temperature
damping parameter in time units, e.g. psec for unit metal, for both 'fix nvt' and 'fix npt'.
So, if set thermostatting as:
........
unit metal
velocity all create 600 222
fix heat all npt temp 600 600 0.2 iso 0.0 0.0 1
........
then temperature will ramp to 600k in about 200 timesteps, but my simulation shows the damping
time is not about 200 but 2000 timestep. The 'timestep' command is 0.001 by default.
the problem is not with fix npt/nvt, but with your expectations.
the Tdamp time applies to *small* deviations from the target
temperature. you expect to almost *double* the kinetic energy
of your system, this is completely unrealistic. normal coupling
through a heat bath is transferring kinetic energy much more
slowly. there is a limit to the amount of kinetic energy that a
nose hoover thermostat can transfer per time (and for a good
reason). for initial equilibration and in order to enforce good
equipartitioning of kinetic energy in your system, a choice of
fix nve with fix langevin is much more effective.
in any case, expecting that a system is in thermal equilibrium
with its (simulated) environment after only 200 steps of MD is
very unrealistic. sorry.