I am currently running eABF(adaptive biasing) simulation using NAMD. I need to restrain a few terminal C-H bonds in the system. I tried including the additional definition for bond constraints in the existing colvar file, but the simulation terminates prematurely, giving the error 'Velocity too high for the H atom; simulation has become unstable'.
The system is adequately minimized, and the classical MD simulations run perfectly. Is there any other way to eliminate the error when Colvars come into the picture other than reducing the timestep?
