something wrong in NPT

I am sorry I did not figure out everything clearly.

The temperature for three parts were calculated:

compute T1 part1 temp

compute T2 part2 temp

compute T3 all temp

and then, they were output like this:

thermo_style custom step temp c_T1 c_T2 c_T3

thermo_modify temp T flush yes
thermo 100

Both ‘temp’ and ‘c_T2’ equal to 350K, ‘c_T1’ equals to zero because the atoms were fixed. But ‘c_T3’ equals to 300K.

I supposed that both temp and ‘c_T2’ equal to 300K and ‘c_T3’(temperature of the whole system) lower than 300K.

thanks,

hangyan

2011-09-22

I am sorry I did not figure out everything clearly.

The temperature for three parts were calculated:

compute T1 part1 temp

compute T2 part2 temp

compute T3 all temp

and then, they were output like this:

thermo_style custom step temp c_T1 c_T2 c_T3
thermo_modify temp T flush yes

what is the compute T computing?

thermo 100

Both 'temp' and 'c_T2' equal to 350K, 'c_T1' equals to zero because the
atoms were fixed. But 'c_T3' equals to 300K.

I supposed that both temp and 'c_T2' equal to 300K and 'c_T3'(temperature of
the whole system) lower than 300K.

well, actually, the temperature as it is "seen" but your fix npt
should have been computed through c_2_temp

in any case, it is impossible top debug an input
in tiny pieces. the documentation about it is
extensive and people have used and tested those
fixes a lot, so that it is unlikely that the fix is in
error. if you believe differently, please provide a
complete(!) and easy to run input that reproduces
your problem, so that somebody that knows the
fix n?t infrastructure well, can debug it and determine,
who or what is at fault.

thanks,
    axel.