[ESPResSo-users] dihedral angle, constraint and pressure calculation

[Arash Azari]

Hello
everyone,

A couple
of questions and I highly appreciate your help:

1- I
would like to simulate a rod-like (stiff) polymer and I have tried
using dihedral angle potential.  
I set
the n=1 (number of minima) and the p=PI (phase parameter). Every time
I get the bond broken error; I have tried very long warm up, even 2
million steps with small time steps 0.0005, but a few steps after
warm up with time step 0.001 the bond broken message appears,
otherwise I should continue the simulation with the same time step
for slightly longer simulation before getting the error message. I
tried to increase the bending constant and time step gradually, but
it does not solve the problem.  
I had a
look at the source codes, although I could not understand too much,
and I saw this notice:  

“Note
that usage of dihedrals increases the interaction range of bonded
interactions to 2 times the maximal bond length!” 

 
I cannot
fully understand the reason and the meaning of the above statement.
To prevent bond broken message, we can increase the maximum bond
length a little and on the other hand the dihedral potential
multiplies it by 2.
What is
the correct way of using dihedral potential for rod-like system? I
mean time step, warm up,  settings, and simulation.  
I had a
look at the manual and I found these features there, GAUSSRANDOM and 
GAUSSRANDOMCUT
, only in the manual and nowhere else. Did they
already implemented in the codes?  
I think
one way to simulate rod-like polymer is using the virtual sites, but
honestly, in my opinion, the user manual for virtual sites is poor.
It will be great if anyone could add some examples (which work!) to
this section. I have tried to use virtual sites a number of times,
but I could not (sorry, I cannot remember the details).  

2- When
we do not use the periodic boundary condition, why we still need the
box length? Does the code need the box length for domain
decomposition for parallel run?  

3- I
would like to impose spherical confinement constraint on my system
and I start with a very large sphere, then decreasing the radius of
the sphere gradually. I thought that the sphere should be surrounded
by the simulation box, say the diameter of the sphere should be
smaller than the box length. In this way, the simulation is too slow
because of the large box. If the box length in this case is only used
to decompose the domains for parallel run, may I start with a smaller
box length?  
4-
Another question related to the pressure calculation in the above
system. Does the pressure calculation procedure work in constrained
and confined systems? Especially non rectangular system where we
cannot use the change_volume command.  

5- When
we get the force on the constraint, constraint force command, is it
the force per area or just net force exerted on the constraint, e.g.
boundary walls.  

Thank
you very much,


Best
regards,
Arash