Communication cutoff xxx is shorter than a bond length based estimate of xxx. This may lead to errors

LAMMPS LAMMPS General Discussion
,
1

HI dear users.
i have a problem when studying the behaviour of bonded particle model.
but i got the warning “Communication cutoff 0.001549 is shorter than a bond length based estimate of 0.004915609034564417”. but it is really weird, since the bond length should be 0.00049 for all bonds.
why i will get a bond length of 0.0049?.
please find attached the image for the reference.
https://drive.google.com/file/d/1mofhOnGUnLPGFmy9PsnM7FwOQ50yv2eg/view?usp=sharing
and then i got the error “Too many neighbor bins”, I checked the boundary length and identified that the boundary length decrease at very beginining but suddently significantly increase to a very large value and then the simulation fail. any reason for this problem.
i got a solution, that when I significantly decreased the timestep to like e-10. this simulaton works and but the speed is really slow. is this means that i just have to use such a slow velocity so that i can maintain stable?

thanks you everyone

2

The bond length you are referring to is the “equilibrium length”. You are discounting that bonds may be stretched. The underlying heuristics that leads to the warning is more complex than what you may think and that the result is close to the the equilibrium length is a bit of a coincidence.

The simulation timestep has to be small enough so that you can have a stable numerical time integration. That depends on the strength of the force constant and the mass of the atoms. If you have very stiff forces, you may be forced to use a short timestep.

3

Hi akohlmey
Many thanks for your response and help. it is helpful.
based on your suggestion, i siginificantly increase the diameter of the sphere. now, the critical timestep did not need to be such small.

4

Adding to what Axel said, for BPM bond styles they calculate the bond length estimate by finding the ~longest equilibrium length and multiplying it by the maximum stretch before the bond breaks (specified in the bond coefficients). I don’t follow what you mean by increasing the diameter of the sphere, but you can increase the communication cutoff using the comm_modify cutoff command to ensure there’s no risk that a bond is lost before it breaks and avoid that warning.

5

I used the FENE bond style,and encountered the same problem. WARNING: Communication cutoff 1.5224620483093698 is shorter than a bond length based estimate of 1.855. This may lead to errors.
Why is the estimated bond length so large?

6

Why should it be shorter?

7

for FENE bond style,i have defined a maximum bond length of 1.5

8

That is irrelevant for the algorithm that is applied to make the estimate.

The fene bond styles reports 0.97*sigma as equilibrium bond length and that is multiplied with a factor to get the estimate.

Since this message is just a warning, you are free to ignore it if you feel the warning is not applicable in your case.

9

ok,thanks for your response.