Tdamp values and energy

LAMMPS LAMMPS Mailing List Mirror
1

Hi,
      I am getting change in cohesive energy as I change Tdamp value of NVT fix.
Tdamp En. (eV/atom) : I have 6750 atoms in the system
0.01 -5.316
0.1 -5.465
1.0 -5.470
10.0 -5.473

I checked the variation in energy and temperature and all seems very well converged. I am unable to understand why there is so much variation. Here is my input script.

2

Hi,
I am getting change in cohesive energy as I change Tdamp value of
NVT fix.
Tdamp En. (eV/atom) : I have 6750 atoms in the system
0.01 -5.316
0.1 -5.465
1.0 -5.470
10.0 -5.473

I checked the variation in energy and temperature and all seems very
well converged. I am unable to understand why there is so much
variation. Here is my input script.

the temperature control interferes with the dynamics
of your system. thus unless you need to raise or
lower the temperature in your system, you want
Tdamp to be as large as possible (i.e. as little
interference as possible) without having a drift
in temperature. with that in mind, your numbers
seem to be converging to a consistent limit.

while checking for convergence with simulation
parameters, you should also do some tests on
how much the system size impacts the results.

axel.

3

Thanks for the help and I am checking system size effect on the cohesive energy. But the issue is, -5.316 is the correct value (got it using DYNAMO) but it is not converging to the value for higher Tdamp values.

Regards
Priyank

4

Thanks for the help and I am checking system size effect on the cohesive
energy. But the issue is, -5.316 is the correct value (got it using DYNAMO)
but it is not converging to the value for higher Tdamp values.

in simulations, nothing is ever absolutely correct.
so what makes you think that dynamo is more correct
than lammps? it could be the other way around.
there are many adjustable parameters, that can affect
results in sometimes subtle ways.

i'm not saying, that i know it better than you,
i am just trying to make a point. :wink:

cheers,
     axel.

5

Are you using metal units? If so, a Tdamp of 10
is very large, i.e. 10 psec - so you would have
to run a very long time to thermostat your Temp
with that.

Steve