CANARY is the multi-object adaptive optics (MOAO) on-sky demonstrator developed by Durham University and LESIA Observatoire de Paris, in the perspective of the E-ELT. Since 2013, CANARY has been operating with 3 off-axis NGS and 4 off-axis Rayleigh LGS and compensating for one on-axis NGS observed with a near IR camera and the Truth Sensor (TS) for diagnostic purpose. In this paper, we present the tomographic performance of CANARY during the runs in 2013. We propose a detailed analysis of the tomographic error leading to the establishment of the CANARY wave-front error budget. In particular we are able to evaluate the tomographic error for each altitude in the atmosphere for a given reconstructor by modelling a set of one-layer covariance matrices. This tool allows us to understand the tradeoffs to be made in the building of the tomographic reconstructor. We present two methods for the wavefront error budget computation. The DTI one uses input system parameters and open loop WFS slopes to estimate the error in a number of independent terms. The DMTS method directly uses the Truth Sensor measurements to estimate the error. We show a good agreement between the two approaches making us confident in our modelling of the instrument. We derive estimations of the Strehl ratio from the error variance and compare them to the recorded IR image Strehl ratio. We find a good agreement between the two, hence validating our wavefront error analysis. Finally we present an on-sky validation of the tomographic reconstruction using LGS based on GLAO and MOAO data. We also quantify the gain brought by the LGS, comparing results obtained in MOAO with 3 NGS and with or without LGS in the wavefront measurements.

Date Published: 21 July 2014
PDF: 15 pages
Proc. SPIE 9148, Adaptive Optics Systems IV, 91482N (21 July 2014); doi: 10.1117/12.2055820

Published in SPIE Proceedings Vol. 9148:
Adaptive Optics Systems IV
Enrico Marchetti; Laird M. Close; Jean-Pierre Véran, Editor(s)