Accurate measures of double stars require accurate calibration of the instrument. Here we present a list of 50 pairs, that are quasi-evenly spaced over the southern sky, and that have Separations and Position Angles accurate at the milli-arcsec, and milli-degree level. These wide angle pairs are suggested as calibration pairs for lucky imaging observations.

We compared several techniques for calibrating angular separation between wide (>1 arcsec) pairs. These techniques are (i) reference pair calibration using {\alpha} Cen AB orbital parameters, (ii) the video drift method, and (iii) the utilisation of an aperture diffraction grating with red filters of different passbands. Separations of 62 pairs were determined using these 3 calibration techniques and compared. It was found that {\alpha} Cen AB and video drift methods are in good agreement. The use of the grating and filter (by measuring fringe spacing) proved unsatisfactory for the broad-band filters, and the use of a narrow band H{\alpha} filter with the grating, resulted in image scales that differed from those obtained using {\alpha} Cen AB reference pair calibration and the video drift method by 0.024 and 0.031 pixel/arcsec (px/arcsec) respectively. A more complete modelling of Fraunhofer diffraction of the H{\alpha} filter and grating produced a difference in image scale of 0.009 px/arcsec. A bias in the diffraction grating method of ~0.1% in the separation of pairs also revealed itself and could not be accounted for. We conclude that calibration against a known pair for which the separation and PA is known with high precision is probably the simplest and best way to undertake image scale calibration.

We report lucky imaging observations of 62 pairs at mid-southern declinations sourced from the WDS with separations larger than 4 arc seconds and magnitude less than 10. The measures comprise separations and PA calibrated against Alpha Centauri AB and drift scans, presented as weighted means of these two calibration methods, with formal internal uncertainties \delta\rho = 80 mas and \delta PA = 0.056(deg). We also compare our measures against 1) extrapolated historic measures, 2) GAIA DR2 data and 3) measures determined from HIPPARCOS and GAIA observations. Our best estimate of our bias against these 3 databases are \rho \approx 10 \pm 30 mas and PA \approx 0.04 \pm 0.08(deg). These formal uncertainties are consistent with the internal uncertainties of \delta\rho = 80 mas and \delta PA = 0.056(deg). We also report Rectilinear Elements for 61 pairs, Grade 5 Orbital Elements for 5 pairs and suggest 5 pairs as optical doubles (4 of which are new).

We present measures for 10 pairs in the constellation of Scorpius using a C14 telescope, Lucky Imaging, and the Reduc software. The separations of Alpha Centauri AB, as determined from the orbital elements of Pourbaix and Boffin (2016), were used as an image scale and position angle calibrator. Our internal uncertainties are ~0.06 arcsec in rho and ~0.06 degree in PA. There is excellent agreement with historic data extrapolated to epoch of observation (~2018.53), and micro-arcsecond positions from the GAIA database where the differences are ~0.05 arcsec in rho and ~0.15 degrees in PA. In addition, we present rectilinear elements for the 10 Sco pairs and orbital elements for two of them. Ephemera are given for these pairs based on both the rectilinear elements and the orbital elements.