Extreme accuracy of the autocollimator-null-indicator

Improving the accuracy of measurements in dynamic goniometry is inextricably linked with achieving the ultimate accuracy of the optical null indicator. This device is part of angle measuring devices operating on the principle of a dynamic goniometer. The autocollimator-null-indicator, when its optical axis coincides with the normal to the controlled reflecting surface, generates an electric pulse, which is used to read the readings of the angular scale of the dynamic goniometer. The ultimate accuracy of the autocollimator-null-indicator is determined by its random error. The random error of measurements of the constant angular position of a reflecting mirror at different times of the day was experimentally studied. It is shown that external noises prevail during daytime measurements – various vibrations, air and external lighting fluctuations, which depend on the distance between the autocollimator-null-indicator and the mirror, and during nighttime measurements the influence of external effects on the measurement results is minimal, which made it possible to obtain the measurement error at night at the flicker noise level. The experimental data are analyzed using the methods of mathematical statistics, Allan variance and wavelet analysis. It is determined that the arrays of random variables characterizing the random error are non-stationary. The minimum value of the random error of the autocollimator-null-indicator was 0.001″.The presented results are of interest to specialists developing and using optoelectronic devices based on the autocollimator.

autocollimator-null-indicator, random error, wavelet analysis, Allan variance