The Colibri Telescope Array for KBO Detection through Serendipitous Stellar Occultations: a Technical Description

TL;DR

The paper describes the design, construction, and testing of the Colibri telescope array, which monitors stars for occultations by trans-Neptunian objects to detect and analyze small distant objects in the solar system.

Contribution

It introduces a novel telescope array setup optimized for serendipitous stellar occultation detection of TNOs, including technical specifications and performance evaluation.

Findings

System performs well under moderate site conditions.

Limiting magnitude at 40 fps is about 12.1 in Gaia G band.

Astrometric errors are typically less than 0.8 arcseconds.

Abstract

We present the technical design, construction and testing of the Colibri telescope array at Elginfield Observatory near London, Ontario, Canada. Three 50-cm telescopes are arranged in a triangular array and are separated by 110-160 metres. During operation, they will monitor field stars at the intersections of the ecliptic and galactic plane for serendipitous stellar occultations (SSOs) by trans-Neptunian objects (TNOs). At a frame rate of 40 frames per second (fps), Fresnel diffraction in the occultation light curve can be resolved and, with coincident detections, be used to estimate basic properties of the occulting object. Using off-the-shelf components, the Colibri system streams imagery to disk at a rate of 1.5 GB/s for next-day processing by a custom occultation detection pipeline. The imaging system has been tested and is found to perform well, given the moderate site conditions. Limiting magnitudes at 40 fps are found to be about 12.1 (temporal SNR=5, visible light Gaia G band) with time-series standard deviations ranging from about 0.035 mag to >0.2 mag. SNR is observed to decrease linearly with magnitude for stars fainter than about G = 9.5 mag. Brighter than this limit, SNR is constant, suggesting that atmospheric scintillation is the dominant noise source. Astrometric solutions show errors typically less than approximately 0.3 pixels (0.8 arc seconds) without a need for high-order corrections.