Tracking errors

The telescope encoders return values in the azimuth-elevation system, so the TRACKING_AZ values from the IMBFITS tables must be multiplied by $cos(\mathrm{el})$ in order to have a reliable idea of the telescope behavior on the sky. The Figure  shows^3.8 three real examples for three different scans. While the tracking errors seem comparable in both directions at low elevation, the dispersion in azimuth is completely flatenned near the pole: the same jittering of the telescope along the azimuth axis results in different jittering on sky, depending on the elevation ( $cos(86^{\circ}) \simeq 0.07$). Question: no wind effect near zenith??

Figure: Antenna position (blue) and tracking errors (red) of three drift subscans from three different on-the-fly scans, with two zoom factors for each. The tracking errors are displayed from a unique central position in order to visualize the dispersion in both directions, and a magnification factor is applied for better visibility. All plots show the azimuth-elevation sphere under an arbitrary orientation and the associated grid. Top: On-the-fly scan from the MRTCAL test data, at about $40^{\circ}$ elevation. Center: On-the-fly scan near the zenith ($86^{\circ}$); subscan drifts along right-ascension, nearly equivalent to elevation scanning for the source and date-time of observation. Bottom: Same, but scanning along declination, nearly equivalent to azimuth scanning.