[lammps-users] drag factor in Nose-Hoover thermostat

Hi Lammps users,

I’ve been playing around with some simple canonical ensemble simulations using the Nose-Hoover thermostat (NH).

Does anyone know what exactly the “drag” factor does to the NH equations of motion? Such a quantity is not discussed in the original Nose and Hoovers articles, and there is not much of a description of it in the lammps documentation. Does anyone know of a reference that introduces drag factor in NH thermostats?


The "drag" is an ad-hoc add-on that is not legitimate NH. It
just helps avoid big oscillations if you change the target T dramatically.
Don't use a drag and you get true NH.