Sources for constellation errors in modulated dispersion interferometers

Dispersion interferometry (DI) is being employed on an increasing number of fusion experiments to measure the plasma density with a minimal sensitivity to vibrations. DIs employed in high-density experiments use phase modulation techniques up to several hundred kilohertz to enable quadrature detection and to be unaffected by variations of the signal amplitude. However, the evaluation of the temporal interferogram can be a significant source for phase errors and does not have an established processing method. There are two non-approximation-based methods currently in use: one using the ratio of amplitudes in the signal's Fourier spectrum and the other using its sectioned integration. Previously, the methods could not be used simultaneously since they differ in their respective calibration point. In this paper, we present a technique to use both phase evaluation methods simultaneously using quadrature correction methods. A comparison of their strengths and weaknesses is presented based on identical measurements indicating one to be more reliable in a more static measurement scenario, while the other excels in highly dynamic ones. Several comparative experiments are presented, which identify a significant error source in the phase measurement induced by polarization rotation. Since the same effect may be induced by Faraday rotation, the results may have direct consequence on the design of the ITER dispersion interferometer/polarimeter as well as the European DEMO's interferometer concept.