Error analysis of Rijnhuizen tokamak project Thomson scattering data

At the Rijnhuizen tokamak project a double pulse multiposition Thomson scattering diagnostic has been operational since 1996. It has been installed for the study of small scale structures in electron temperature (T-e) and density (n(e)). This diagnostic measures T-e and n(e) with high spatial resolution (3 mm full width at half maximum, i. e., 2% of the minor radius) and high accuracy (3%-4% of T-e and 2%-3% of n(e) in the range of 50 eV-6 keV and n(e) 5 x 10(19) m(-3).) In this article an extensive error analysis is performed on both statistical and systematic deviations. It is found that the instrument function of the detection branch has a smoothing effect on the noise. This reduces the statistical error on the T-e and n(e) measurements on each spatial position, because the resolution of the instrument is oversampled. The long tail of the instrument profile of the entire diagnostic has a significant effect on the systematic deviations in the T-e and n(e) determination. However, it does not affect the relative size of the small scale structures on T-e and n(e), and for this this reason does not hamper the study of these structures. (C) 1999 American Institute of Physics. [S0034-6748(99)02704-5].