Observations of 20 open clusters at the Maidanak Observatory were performed using the Zeiss-1000 telescope in the winter of 2018. As the light receiver, a CCD camera Apogee Alta U9000 (3K x 3K, FOV 19.5'x19.5') was used. All observations were performed using BVRI light filters. The exposure time, depending on the filter was from 60 to 150 seconds. Due to the instability of the telescope, images of the stars are slightly elongated. For each observation night, calibration images were obtained. Astrometric processing was carried out after filtering hot pixels. For testing purposes, we used two methods of reduction: by the Astrometrica program and the author's set of programs with primary processing in the MIDAS environment. The reduction of 280 frames was performed using a polynomial of the third degree. As reference stars in different processing options, the Tycho2, UCAC4 and Gaia DR2 catalogs were used. Due to the small field and the small number of reference stars, the processing in Tycho2 showed the worst result. The results obtained by author programs with reference catalogs UCAC4 and Gaia, showed good convergence. Integrated catalog of 18 thousand stars upto 18.5 m was obtained. The accuracy of the catalog is σ RA =0.011" and σ DEC =0.013", the accuracy of single observation on RA (0.030-0.032)" and DEC (0.035-0.037)". At the same time, the accuracy of intraframe processing with the Gaia reference catalog turned out to be 2-3 times better than the analogous one with the UCAC4 catalog. The accuracy of processing elongated star images by the Astrometrica program is on the average 3-4 times worse than by the author's programs. Also a systematic dependence of the pixel scale on the using filter was found. Based on the results of the test photometric processing of NGC869 cluster a photometric estimate of the accuracy of the observations was obtained. The internal errors of one measurement of the stellar magnitude in the instrumental system bvr have the values σ bvr = (0.03-0.05) m. After the reduction of instrumental stellar magnitudes to the Johnson BVR system, the errors of one measurement are σ BVR = (0.05-0.08) m. The processing of observations showed the possibility of carrying out differential observations on the telescope Zeiss-1000 with an accuracy of one observation, not worse than 0.04", and near 0.02'' after improving telescope management.