# regarding Tdamp in fix nvt

Hi,

I have some questions about the 'Tdamp' factor in 'fix nvt'. I am
simulating a simple WCA system of 10,000 particles in constant NVT
ensemble in 2 dimensions using 'fix nvt'. I want to know what is the
appropriate value of Tdamp for such a system? If deltat=0.0001, then how
many MD-steps should it take to relax the temperature for Tdamp=10.0? To
obtain the appropriate Tdamp value I ran 3 simulations at the same
temperature=0.5 for Tdamp=1.0, 10.0, 100.0 each for 9,00,00,000 steps and
saw the temperature profiles. In all the cases temperature fluctuates from
0.485 to 0.515 which seems to me very large for a system containing 10,000
particles. Is such a large temperature fluctuation is normal with 'fix
nvt'?

Chandana

Hi,

I have some questions about the 'Tdamp' factor in 'fix nvt'. I am
simulating a simple WCA system of 10,000 particles in constant NVT
ensemble in 2 dimensions using 'fix nvt'. I want to know what is the
appropriate value of Tdamp for such a system? If deltat=0.0001, then how
many MD-steps should it take to relax the temperature for Tdamp=10.0? To
obtain the appropriate Tdamp value I ran 3 simulations at the same
temperature=0.5 for Tdamp=1.0, 10.0, 100.0 each for 9,00,00,000 steps and
saw the temperature profiles. In all the cases temperature fluctuates from
0.485 to 0.515 which seems to me very large for a system containing 10,000
particles. Is such a large temperature fluctuation is normal with 'fix
nvt'?

the size of fluctuations of the instantaneous temperature
does not only depend on the number of particles, but also
on how compressible your compound or rather its potential is.

you can easily determine the typical fluctuations from a run
starting from a well equilibrated structure and then continued
with fix nve. unless you want to study non-equilibrium phenomena,
you want the coupling from fix nvt to be very weak, so that it
doesn't impact the dynamics of your system more than absolutely
necessary to model the impact of the coupling to the environment.

cheers,
axel.

If deltat=0.0001, then how
many MD-steps should it take to relax the temperature for Tdamp=10.0?

10.0/0.0001 = 100K steps.
See the fix nvt doc page.

Steve