As a follow-up to our previous effort to consider the effects of steel lamination core loss on 3-D transient magnetic fields, a new algorithm is proposed to eliminate possible convergence problems and improve computation efficiency when a strong eddy core-loss component is involved. The proposed approach is validated by two application examples in terms of the power-balance testing, that is, the increase of input power between considering the core-loss effects and without considering core-loss effects must be equal to the total core losses.