Study On The Five-bar Mechanism Measuring Instrument And Its Calibration Methods

The five-bar mechanism measuring instrument is a kind of parallel mechanism used to measure the planar curves.It’s lightweight, flexible and with high repetition accuracy.But the difficulty in parameter calibration is needed to be solved. Research of error analysis and calibration in five-bar mechanism measuring instrument was conducted as follows.First, this paper introduces the mechanical structure of five-bar mechanism measuring instrument. The measuring instrument was redesigned for the smaller radiu subway wheels.The forward and reverse solution model for measuring instrument was wrote out and the measuring filed of the instrument was simulated.Secondly, the analysis of the error sources of the instrument was carried out.The parameter error and quantization error were simulated using Matlab.Results are as follows: error caused by the bar length error and encoder position error of parameter error is about 10-3mm magnitude;the quantization error and error caused by the error in the zero position of encoder(3600 lines) is in ±0.03mm which has the greatest impact. By compensating for half a pulse can basically eliminate the overall trend of the quantization error. The transfer function between parameters error and the end position error was deduced using a closed-loop vector chain method and the error model of measuring instrument was estanlished.According to the error model, sensitivity coefficient of every parameter was calculated. The results showed that the sensitivity coefficient of the encoder zero position error is the highest.Then the principle of the measuring wheel radius compensation and the two-point method of traditional radius compensation method were analyzed..Fourier series was adopted to fit the coordinates which was obtained by offset and quantization JM3 model standard curve to compensate the radiu.The result shows that this method improved significantly compared with the two-point method. The error using derivative of fitting curve method is in±3×104mm.Finally, four sets of calibration simulation was set on the path straight horizontal and vertical line,on which the calibration models were established based on the error model established before.The four tests are:1)consider the parameter error, does not consider the quantization error;2) consider the parameter error and quantization error;3)consider the encoder zero position error and quantization error;4) the error of rod length and encoder position is equivalent to the error of the zero position of encoder on the basis of test two, design the orthogonal experiment to avoid particularity.The simulation results show that when the coefficient matrix of calibration equation contains all the parameters components, the matrix was so morbid as to small error will cause huge change in the solution of the equation. When the coefficient matrix only contains the encoder zero position components,matrix is very stable. The accuracy of the parameters calibration of the horizontal line is better than the vertical line in most test. When the precision of rod length and encoder position is more higher,the method in test 4) can be more significantly reduce the error in calibration.