The energetic cost of burying charged groups in the hydrophobic core of lipid bilayers has been controversial, with simulations giving higher estimates than certain experiments. Implicit membrane approaches are usually deemed too simplistic for this problem. Here we challenge this view. The free energy of transfer of amino acid side chains from water to the membrane center predicted by IMM1 is reasonably close to all-atom free energy calculations. The shape of the free energy profile, however, for the charged side chains needs to be modified to reflect the all-atom simulation findings (IMM1-LF). Membrane thinning is treated by combining simulations at different membrane widths with an estimate of membrane deformation free energy from elasticity theory. This approach is first tested on the voltage sensor and the isolated S4 helix of potassium channels. The voltage sensor is stably inserted in a transmembrane orientation for both the original and the modified model. The transmembrane orientation of the isolated S4 helix is unstable in the original model, but a stable local minimum in IMM1-LF, slightly higher in energy than the interfacial orientation. Peptide translocation is addressed by mapping the effective energy of the peptide as a function of vertical position and tilt angle, which allows identification of minimum energy pathways and transition states. The barriers computed for the S4 helix and other experimentally studied peptides are low enough for an observable rate. Thus, computational results and experimental studies on the membrane burial of peptide charged groups appear to be consistent.