MD at constant temperature in metals

[call_me_Al]

Hi VASPers,

I am trying to run some MD simulations and take "snapshots" of typical ion positions etc during a run at a constant temperature. My INCAR typically looks like this:

ISMEAR = -1
SIGMA = 1
NBANDS = 96
ALGO = Very Fast #exact
MDALGO=1
IBRION = 0 ; NSW = 500 ; NWRITE = 0; LCHARG = .FALSE.
TEBEG = 25000; TEEND = 25000
POTIM = 1
NELMIN = 4

It seems to run and give me different ion positions in the CONTCAR, but I notice the temperature is decreasing at every timestep. I would like to keep a constant temperature and volume during the run.

for example, the first and last timestep energy info look like this:
1 T= 23496. E= -.80540227E+01 F= -.17165405E+02 E0= -.14500394E+02 EK= 0.91114E+01 SP= 0.00E+00 SK= 0.00E+00
500 T= 12219. E= -.84088553E+01 F= -.13146984E+02 E0= -.10471487E+02 EK= 0.47381E+01 SP= 0.00E+00 SK= 0.00E+00


does anyone have ideas about how to keep the temperature constant?

thanks!

[call_me_Al]

I just wanted to add that I am only using four atoms in the POSCAR file due to computational restrictions,

could this be a source of large fluctuations etc?

I ran a 2000 timestep 1 fs cell and it converged on a lower temperature (17 kK insted of the INCAR 25 kK)