Solution principles for a new generation of precision self-correcting multi-axis machines

Based on the premise that machinery cannot be built perfectly, this paper formulates plausible solutions for the active correction of the inherent inaccuracies of "less-than-perfect" machines in order to achieve or exceed the performance of precision machinery built by conventional methods which typically aim to build a "perfect machine". Three problem areas central to the attainment of active or self-correcting capabilities are addressed. First of all, error correcting strategies based on mathematical models of error propagation are addressed through the development of a generalized first order volumetric error model of a multi-axis machine. Second, examples of real-time in-process error sensing methods are analyzed. Finally, strategies based on adaptive and conventional control approaches by which corrections can be fed back and executed by the machine's controller are introduced. For all three elements above, examples of developed systems are presented. It has been shown that the development of self-correcting machines is within the realm of current technology and that significant improvements in precision can be achieved at a relatively modest cost.