How can I fix this error:
PPPMDipole dipole Newton solver failed, using old method to estimate g_ewald.
or Ewald/disp Newton solver failed, using old method to estimate g_ewald.
I got this warning which causes an error which results in missing bond atoms.
I chose a different cutoff value but it didn’t help.
These are warnings not errors. They are usually an indication of inconsistent per-atom data.
They can be easily avoided by manually setting g_ewald to a meaningful value instead of letting LAMMPS estimate it. Of course, silencing the warning will not really resolve any underlying problems with the system.
The warning usually is a result of an unsuitable geometry or per-atom parameters. Changing the cutoff is not likely to help with that.