The distributed cavity phase (DCP) shift is a significant source of uncertainty in the accuracy budget of several atomic fountains (see for instance [1 2]). This effect arises when the moving cold atom cloud interacts with the imperfectly stationary microwave field inside the Ramsey cavity. The effect depends on several parameters: The cavity geometry which determines the phase distribution in the cavity, the atomic cloud position and velocity distributions, the microwave power, the overall fountain geometry and the detection parameters which modify the effective velocity distribution of detected atoms. Due to this large number of parameters, there is no straightforward method to measure the distributed cavity phase shift in atomic fountains. So far, the related uncertainty has typically been estimated based on one of several available models for the phase distribution [3 6], on measurements of the atomic cloud parameters and their stability in time and on worst case calculations of the clock shift.