Hi:
I am working on a simple LJ simulation. The molecules are just two beads with a charge on them.
I have been running with an nvt of
nvt temp 1.0 1.0 100.0
My damping parameter is 100 taken from the docs.
However when running with that parameter my temperature coming out in the thermo output is quite a bit larger, say 2.
When I go with at damping parameter of 1.0, I get my thermo output temp to be essentially 1.0, which is great.
It seems clear to me that i want a damp of 1.0 in my case, based upon my thermo temperature output.
Am I thinking right here?
What is the purpose of the damp? Under what conditions would I need to set the damp larger?
The 100 fro the doc page is 100 timesteps, not just “100”. Assuming LJ units, the default timestep is 0.005, so that the suggested damping should be 0.5. This is why your setting of 1.0 gives better results.
Damping defines the time constant for the thermostat. Please consult a MD textbook for more details.
Ray
The 100 fro the doc page is 100 timesteps, not just "100". Assuming LJ
correction. it is *time units* and not time steps.
units, the default timestep is 0.005, so that the suggested damping should
be 0.5. This is why your setting of 1.0 gives better results.
thus a damping of 100 time units corresponds to 20000 time steps.
with 1.0 you reduce the time constant until the thermostat will
attempt to reach the desired temperature by a factor of 100.
Damping defines the time constant for the thermostat. Please consult a MD
textbook for more details.
and particularly have a look at the discussion of equilibration and
equipartitioning of kinetic energy.
axel.