Research On The Error Suppression Of All-digital Resolver-to-digital Converter Based On The Delayed Signal Cancellation

Resolver to digital converter is used to obtain the angular position informationof the rotor from two outputs of the resolver. The traditional method, which uses aspecial integrated RDC chip, has the relatively high cost and cannot inhibit thesignal errors (such as DC offset, amplitude error, phase error, harmonic error andnoise interference). However, the all-digital resolver-to-digital converter (RDC)with ability of elimination of signal errors has been more and more widely used dueto its low cost.This paper firstly detailedly introduces the whole structure and the principle ofthe resolver angle measuring system. The demodulation algorithm is improved tosuppress the delay of the excitation and noise interference, which otherwise willweaken the measurement performance. The influence of different kinds of signalerrors to angular position measurement is deeply analyzed, which provides atheoretical basis for the new proposed all-digital RDC algorithm.Secondly, the paper introduces the expression of resolver signals from the pointof space vector, and then analyzes the principle of the cascaded delayed signalcancellation (CDSC) which can eliminate different kinds of signal errors.Considering the issues of rapid response in phase-locked loop, this paper analyzesthe structure of the delayed signal cancellation (DSC) operation unit separately inrotary and static coordinate systems, and improves the proposed algorithm. Afterthat, the relatively harmonic gain is introduced to analyze the influence of thediscretization error during the digital implementation. In order to inhibit this error,linear interpolation method is used to improve the proposed algorithm. Theimproved algorithm is structurally simple and easy to implement based on theresults of the simulation and experiment.Thirdly, in order to improve the low-response of the frequency detection inCDSC, this paper proposes a new method combined with the adaptive notch filter(ANF) to detect the frequency of the input signals. This method is further improvedto satisfy the condition that it can simultaneously receives two input signals andaccurately measures the unique signal frequency and the rate of change of frequency.Through the analysis of frequency detection simulation, the proposed detectionmethod is proved to have the ability of rapidly and accurately tracking the signalfrequency, robust performance with respect to distortions of the input signal, and therobustness with respect to uncertainties and variation of its internal parameters.Finally, the DSP-based digital RDC system is built to verify the effectivenessof the proposed algorithm. The hardware circuit of the special integrated RDC chipis added in order to compare its results of angular position measurement with that of all-digital RDC for further error analysis. This paper mainly introduces the softwaredesign in digital filter and further error suppression. The experiments on signaldemodulation, CDSC, and different states of rotor motion are conducted to validatethe correctness of the proposed methods. The experimental results show that theCDSC combined with the ANF technology can effectively suppress the differentkinds of signal errors to accurately obtain the angular position of the rotor.