[lammps-users] A question of Tdamp for nvt at metal units

Dear lammps users,

I want to do simulation with a constant temperature.
So now I use npt command as following:

fix 1 all nvt 500.0 500.0 1.0

In this simulation, unit is metal.
Also as I know that Tdamp value means that the relaxation (damping) process explains how quickly oscillations get smoothed out by the NVT integrator (the smaller the constant, the “sharper” the temperature corrections), so the value of this parameter has nothing to do with how quickly the temperature adjusts-only with how quickly deviations from the linear temperature progression are “wiped out”.

For my simulation, even if I use nvt or npt, I think Tdamp should be same value.
For a simulation with constant temperature in metal units, is 1.0 value of Tdamp is reasonable?

Thank you in advance.

Choi, HeeSung

The units of Tdamp are time, which for metal units is picosec.
So, yes, a value of 1 psec is reasonable, for either nvt or npt.
I usually make Pdamp (for NPT) about 10x larger than Tdamp.
So 0.1 psec for Tdamp and 1.0 for Pdamp is also fine.

If your system is not at T_target, then Tdamp is also the
timescale on which the system will adjust to become T_target.


2009/10/21 ChoiHeeSung <[email protected]...>: