Linear Motion Systems. A Modular Approach for Improved Straightness Performance

Abstract

This thesis deals with straight motion systems. A modular approach has been applied in order to find ways to improve the performance. The main performance parameters that are considered are position accuracy, repeatability and, to a lesser extent, cost. Because of the increasing requirements to positioning systems, concerning accuracy, repeatability and velocity, it becomes more and more difficult to meet these requirements when using merely mechanical means. The errors in position occurring as a result of e.g. external forces, wear or thermal expansion, are the reason that open-loop straight motion systems can only operate in the sub-micrometer range when very strict conditions are applied. When the required accuracy becomes smaller than the manufacturing tolerances achievable, active compensation of errors becomes indispensable. On the other hand, by the application of active error compensation, the requirements to the manufacturing accuracy can decrease. By means of this feedback of straightness errors, the problems and bottlenecks shift from the mechanical domain towards the electronics and control engineering domain, where they are better solvable. The eventual performance of a feedback controlled system is not limited any more by fabrication tolerances, bearing properties, thermal influences or external loads, but is mainly limited by the resolution of the sensor applied, and by the dynamical behaviour of the mechanical system. When aiming at a position accuracy in the sub-micrometer range, also vibration problems become important. Therefor, a completely different approach has been applied. Instead of aiming for the highest possible open-loop stiffness of the positioning system relative to the ground, an actuator stiffness equal to zero is proposed. The zero stiffness bearing results in a decoupling of the position, while preserving the force coupling to the world. This means that the position of such an object (e.g. a specimen holder), supported by a zero stiffness bearing, may be coupled to the position of another object (e.g. a lens on which no external forces are allowed) by means of a control loop, resulting in a virtual stiffness relative to each other. The forces, needed for positioning the moving object, cause a reaction force on the world, while the vibrations (position errors) of the world do not influence the positioning accuracy of that object. This concept offers possibilities in the design of precision positioning systems, where otherwise vibrations from the outer world would have a too large disturbing influence. Within the scope of this research on zero stiffness actuators, a permanent magnetic bearing element has been designed and built, which is able to generate a static bearing force that is without mechanical contact, without stiffness and without power consumption. The permanent magnetic zero stiffness bearing principle has been patented in the Netherlands; an international patent application is pending.