Fixed electric charge is believed to play an important role in the transfer of charged solutes in nanofiltration. Electrochemical measurements are useful tools for the determination of electrochemical perm-selectivity, which can be related to the fixed charge density by using a model. In the case of monolayer membranes, the electrochemical perm-selectivity is determined, in particular, through the measurements of stationary membrane potential. With composite/asymmetric NF membranes, the interpretation of this measurement is complicated by the membrane multi-layer structure. In the concentration-step technique, a membrane is equilibrated with an electrolyte solution and left in contact with this solution from the support side alone. The active membrane surface is suddenly touched by a pendant drop of solution of a different concentration, and the electrical response to this is tracked with a pair of reversible electrodes. In the very first moments after the touch, the whole concentration difference is located within the membrane active layer, and the initial electrical response is controlled by its electrochemical perm-selectivity. The characteristic time of relaxation of transient membrane potential is governed by the diffusion permeability of active layer and the porosity of and diffusivity in the membrane support. The concentration-step technique was used to study the electrochemical properties of active layers of a commercial polymer NF membrane (PES 10) and a laboratory-made nano-porous ceramic membrane (provided by University of Twente) in KCl solutions of various concentrations. At pH 6, the fixed charge of polymer membrane was found to be negative whereas that of ceramic membrane was positive. The concentration of fixed charges ranged from 0.03 kmol/m(3) to 0.17 kmol/m(3) depending on the salt concentration. In contrast to the previous study [A.E. Yaroshchuk, A.L. Makovetskiy, Yu.P. Boiko and E.W. Galinker, J. Membr. Sci., 172 (2000) 203-221], the data acquisition was fully computerized, which enabled LIS to carry Out an extensive statistical analysis. Due to that, in particular, we could identify statistically significant deviations from the theoretical model at very short times. The principal reason for these deviations probably is the increase in the contact area immediately after the touch, which is not accounted for in the theoretical model. The poor reproducibility at short times appears to be caused by the variability of membrane local wetting state caused by the Inhomogeneity of hydrophilicity of its surface. By using separately available salt rejection data and the characteristic relaxation times estimated from the concentration-step measurements, we could also estimate the effective diffusion coefficient in the membrane Support layers.