Item

Abstract

The vibrations affecting the telescopes and downstream optics largely contribute to the optical path length error in interferometry and constitute an important limitation for fringe tracking sensitivity and high-contrast imaging. At the VLTI-UT, vibrations are measured and compensated with the MANHATTAN-II system. This system currently measures the acceleration in the first three mirrors of the Coud´e train with piezo-electric accelerometers and sends the compensation error to the delay lines. The GRAVITY+ program and the next generation of the VLTI planet-finding instruments require the current Optical Path Distance (OPD) residual of approximately 200 nm rms in the UTs to be smaller than 100 nm, and consequently an extension of the MANHATTAN-II system is envisaged to all the Coud´e train mirrors that are known to be affected by high-frequency vibrations. In this work, we present the planned hardware upgrade of the Manhattan system and an analysis of the self-intrinsic noise of MANHATTAN-II. Piezo accelerometer noise depends on parasitic electric charges, generated for example from long cables. The noise theoretical model is experimentally tested with a dummy accelerometer reproducing the equivalent circuit, and by a three-sensors coherence analysis. We find that the intrinsic-noise for M2-UT1 is of the order of 2 μg/Hz^0.5 in the amplifier bandwidth 1 Hz – 25 kHz. Accounting for all sensors in every mirrors in the extended configuration, the Manhattan-II self-intrinsic noise corresponds to an overall equivalent OPD rms of 166 nm for the shortest possible cable lengths configuration, before the fringe tracker. We suggest that sensors with sub- μg/Hz^0.5 noise could be necessary to approach the ultimate limit of the atmospheric piston calculated from a Kolmogorov model.