question about iso keywork in fix npt command

Dear all,

1517383345429.png

Dear all,

The lammps that I am using is 2016Nov17. I met a problem when I use fix
npt command. In the manual, it's said that

*Coupling means two things: the instantaneous stress will be computed as
an average of the corresponding **diagonal components, and the coupled
box dimensions will be changed together in lockstep, meaning coupled **dimensions
will be dilated or contracted by the same percentage every timestep.*

However, when I use *fix 1 all npt temp 1120 1120 0.02
   iso 0.0 0.0 0.2 *

I output lx, ly and lz every timestep and I found that the three
directions are not coupled at all. While x and z dimensions are
contracted, y dimension is dilated, and the percentages are totally
different. Here are the output:

Is there anything wrong with lammps? Or I don't understand the manually
correctly?

​or does your input not do what you think it does?

when i replace fix nve in the melt example with :

fix 1 all npt temp 2.5 2.5 1.0 iso 1.0 1.0 1.0

i do get the expected isotropic cell length change. mind you, this is with
the current version of LAMMPS (17 Jan 2018)

Step Temp PotEng Pxx Pyy Pzz Lx Ly Lz
       0 3 -6.7733681 -3.7486388 -3.652099
-3.7093134 16.795962 16.795962 16.795962
      50 1.5326344 -4.832688 3.7564729 3.5563042
3.555895 17.050199 17.050199 17.050199
     100 1.3180211 -4.6517748 2.1140048 2.1088472
2.1729586 17.428678 17.428678 17.428678
     150 1.2391857 -4.5372736 1.6289071 1.4960543
1.5312826 17.644295 17.644295 17.644295
     200 1.2234367 -4.4643994 1.1631854 1.2228478
1.3365478 17.786172 17.786172 17.786172
     250 1.2131804 -4.3769443 1.1961009 1.1946676
1.1760637 17.877547 17.877547 17.877547

​you have to provide more convincing proof that fix npt is not working as
documented.

axel.​

1517383345429.png