something wrong in NPT

Dear Andrew,

Thanks so much for your suggestions.


Dear Andrew,
Thanks so much for your suggestions.

You're welcome. I'm not sure my suggestions were correct. (See below.)

What's going ong for your 'fix rigid' test? I really appreciated if you can
post those tests and perhaps we could discuss about it.

I have many of these tests. Before posting my input files, I tried to
simplify the system. The problem I was having appears goes away.
(I'm not sure if I should be happy, or sad, or just embarrassed.)

Perhaps it would be helpful for me to post a working example.
I -will- post my most successful example files later today once I feel
more confident about why this didn't work earlier. (I would like to
isolate what went wrong.)

The only way that I know of which will allow you to do what you want
is to add harmonic bonds connecting the atoms you want to immobilise
together at the correct distances. You can also use additional
harmonic angle constraints to further prevent the system ("par1") from
moving. Then run your simulation in the ordinary way integrating the
equations of motion on both parts 1 and 2. ("fix 2 npt all ...").
This is the safest choice.

I'm hoping I am wrong about this. Again, if I get fix rigid to work,
I'll post what I did. My most successful test so far uses fix rigid
along with neigh_modify exclude (par1 par1). I'll provide more
details in my next post.


Because several people have posted questions on this topic, I am
posting an example of a simulation of a solid/liquid interface under
NPT conditions using fix "rigid" to immobilize the solid.

The attached file is a gzipped archive containing several files:
"" "", a README file, and a graphic image that
depicts the geometry of the system. (Please let me know if it does
not get included. Gmail sometimes blocks .gz files.)


I created two groups named "mobile" and "immobile". Both are made up
of identical water molecules.
Here are the critical lines from my lammps input script:

fixrigid_immobile_example.tar.gz (182 KB)

By the way, please disregard the "README.TXT" file which I
accidentally included in my last post.

(I forgot to change that file. It is full of me ranting about
problems I ran into with earlier tests. Please ignore it. Fix rigid
appears to work well with NPT.)

Great - please consider posting this example to the LAMMPS user
scripts page: