High-quality microprocessor system for position, velocity, and acceleration measurements of electric drives

Abstract

531.76.66 Various aspects of the design of microprocessor systems for velocity, acceleration, and position measurements are considered. The velocity-measuring method described differs from known methods in that it has minimal errors in the range of velocities from fractions of a rotation to the maximum rotation frequency and requires a microprocessor. The position and acceleration are measured at the same time as is the velocity.One method of increasing the accuracy of modem electric drives of machine tools and industrial robots is by improving the dynamic and static characteristics of transducers in feedback circuits [1, 2]. The velocity of a controlled drive is generally measured by analog transducers, which reduces the control accuracy, especially in the range of low and ultralow values. Moreover, digital drives (to close the main feedback circuit) as well as analog drives (to close the velocity circuit) are used in autonomous follower drives.Coding transducers with magnetic or optical pickup are used to obtain information about the parameters of motion of an object. Essentially, the velocity is measured by two principal methods [2]: counting pulses from a coding transducer in a fixed interval of time and counting pulses from an auxiliary oscillator during coding transducer pulses. In the first case, the measurement is effective in the range of high velocities and in the second, low velocities. The position of the actuators is determined by directly counting pulses from coding transducers and then covering the count into the actual position.Both methods involve the use of equipment and are simple to apply. The quality of the measurements, however, is not always justified. We consider the salient features of the two methods in greater detail and estimate the data acquisition error.The essence of the direct method consists in measuring the number of sensor pulses in a given period of time. The rotational velocity of the sensor disks is calculated aswhere T is the measuring interval, N is the number of sensor pulses counted in the given time, and Z is the number of pulses per rotation. Even though the characteristic of the given method of measurements is linear over all ranges, measurement of slow velocities is a very complicated process. The main reason why the given method has an error of velocity measurement is that the number N is an integer since most often the pulses are counted by standard counters. As a rule, however, a fractional number of pulses fits in the measuring period T. Accordingly, that period must be increased to reduce the errors. For a wide-range sensor, the necessary measuring period is calculated for the minimum velocity since the sensor pulse frequency then is also a minimum.We consider the distribution of the error of velocity measurement by the given method, depending on the velocity and the measuring period (from Fig. 1).

DSP algorithm for the extraction of dynamics parameters in CNC machine tool servomechanisms from an optical incremental encoder

“…Consequently, the algorithm provides better results than previously done research, where the obtained results for the dynamics recuperation were 0.5% in speed [16], and there were even peaks of 5% in jerk [11]. Greater errors are expected in jerk because of the cascade of differences from the position to the jerk, as shown in the algorithmic approach in Fig.…”

“…The responses in the signal reconstruction were improved, but the computational load was high. Although ideal differentiation is not possible, several techniques have been presented such as FIR filter implementation [12] and another focused to estimate low and high speed [16] by transducers and finite-difference implementation.…”

An indirect method of determining an error of the angular velocity of the electric drive based on the reference pulse signal of variable frequency

“…Для высокоточного регулирования ССЭ необходимо удерживать отно-сительную погрешность в пределах ±0,02 % [14]. Из рис.…”

Iterative method of measurement with a given accuracy for angular velocity errors

“…The relative error is needed to hold on within ±0,02% for the highly accurate SPED regulation [8]. Fig.…”