[lammps-users] bug in NVT integrator?

Dear Axel, dear Steve

I think you both are right. At least in the density-profile of the amino-acid I can see that the bio-molecule is kept at a fixed position, maybe because it is prevent from moving by the surrounding frozen water molecules. To make a further check I wanted now to write out the temperatures of water and amino-acid separately, but I have problems with the syntax (see below).

First I defined the two groups, then I set the compute statements defining the temperature variables c_watertemp and c_phetemp which should be output by the thermo_style command. There is always the measure "ERROR: Could not find thermo custom compute ID". So what have I done wrong and how I have to use the thermo_modify for this purpose. I am a bit confused by the description in the tutorial.

group phenylalanine molecule 1
group water type 11 12
compute c_watertemp water temp
compute c_phetemp phenylalanine temp

thermo_style custom step c_watertemp c_phetemp etotal
thermo_modify temp c_watertemp

Thank you.

By the way, can I circumvent the problem with the frozen solvent by separately thermostating water and amino-acid?

Best regards

Steve Plimpton <[email protected]> 05/13/09 3:36 PM >>>

Axel's suggestion is a good one - viz your system.
It could be the classic flying icecube - your system has
frozen and is drifting with a mean velocity. You could
also monitor the center-of-mass drift with fix com.