The DC operation of the retarding field analyzer (RFA) was proposed as a method for directly measuring ion temperature fluctuations in the edge plasmas of fusion devices. The method is based on the relation for the RFA current–voltage (I–V) characteristic resulting from a common RFA model of shifted Maxwellian distribution of the analyzed ions and two-point measurements on the exponentially decaying portion of the I–V characteristic with two differently DC-biased RFA electrodes. In this case, the ion temperature is retrieved from the ratio of the corresponding ion currents. For a 3-grid RFA in configuration, where retarding positive potentials are applied to the second grid and the collector, the method remains applicable as long as the ion-induced secondary electron emission (SEE) yield from RFA electrodes is within the range γi < 0.2. At higher values, the presence of SEE distorts the measurements of ion temperature fluctuations due to the uncompensated dependence on the shift potential. This paper discusses the DC RFA operation in an alternative configuration with retarding positive voltages applied to the first and third grids and ion currents measured from the negatively biased second grid and collector. It is shown that the presence of SEE from RFA electrodes in this configuration does not influence the measurements of ion temperature fluctuations. An important point is the application of effective RFA calibration for ion temperature fluctuations measurements using the mean ion temperature and currents ratio values obtained respectively in standard AC ramping and DC modes. Supported experimental results from the small tokamak ISTTOK are also presented.