The E and B experiment (EBEX) is a balloon-borne telescope designed to measure the polarization of the cosmic microwave background with 8' resolution employing a gondola scanning with speeds of order degree per second. In January 2013, EBEX completed 11 days of observations in a flight over Antarctica covering $\sim$ 6000 square degrees of the sky. The payload is equipped with two redundant star cameras and two sets of three orthogonal gyroscopes to reconstruct the telescope attitude. The EBEX science goals require the pointing to be reconstructed to approximately 10" in the map domain, and in-flight attitude control requires the real time pointing to be accurate to $\sim$ 0.5$^{\circ}$ . The high velocity scan strategy of EBEX coupled to its float altitude only permits the star cameras to take images at scan turnarounds, every $\sim$ 40 seconds, and thus requires the development of a pointing system with low noise gyroscopes and carefully controlled systematic errors. Here we report on the design of the pointing system and on a simulation pipeline developed to understand and minimize the effects of systematic errors. The performance of the system is evaluated using the 2012/2013 flight data, and we show that we achieve a pointing error with RMS=25" on 40 seconds azimuth throws, corresponding to an error of $\sim$ 4.6" in the map domain.
↧