On-machine Metrology for Surface Fabrication

Lead Research Organisation: University College London
Department Name: Physics and Astronomy

Abstract

It has oft been said that, 'if you can't measure it, you can't make it'. Measurement is fundamental to manufacturing technology, and requires specialised and broad knowledge, not only to make a critical measurement, but also to interpret and apply the results appropriately. The old-fashioned way to polish precision lenses and mirrors is to rub them in a controlled manner using a water-slurry of polishing powder with a polishing tool (which might in the later stages be the optician's thumb!). In most applications it is necessary to achieve both an excellent smooth polish free of defects, and a precisely shaped surface (we call it 'surface-form') good to a few millionths of an inch. To achieve this requires many cycles of measurement and polishing. A modern computer controlled polishing machine, such as produced by one of the project-partners Zeeko Ltd, can speed up and control the process, making it more automated and predictable. Nevertheless, repeated cycles of measurement are still needed, because of the underlying complexity at the microscopic level of the physics and chemistry behind polishing. In practice, this usually means de-mounting the lens or mirror from the polishing machine-tool, and moving it to a measurement instrument, which increases production-time and introduces risk of damage. The challenge of measurement becomes acute when trying to manufacture precision surfaces which have complex forms. These include 'aspheres' (surfaces which differ from part of a sphere), and the truly unruly surfaces called 'free-forms' (which may have seemingly random humps and hollows like a Pringle). Today, the technology to polish such complex surfaces in a controlled manner is well ahead of the ability to measure them. It is the measurement part of the cycle which is severely limiting the accuracy that can be achieved, and thwarting the ability of industry to capitalise on the advantages which such surfaces can confer. So why does industry want these complex surfaces? Consider two examples. In optics, complex surfaces provide the designer with more features that can be changed in the computer, when designing a particular lens or mirror. In general, this means that the same job can be done with fewer pieces of glass (lighter, more compact systems), or better performance can be achieved (sharper images). In a completely different field - medicine - artificial knee joints are complex saddle-like forms, and superior quality can increase the joint's life in the patient. To make sense of the increasing need to measure and control complex surfaces requires breadth of knowledge, spanning measurement instrumentation, computer-interfaces, data-analysis, software techniques, sources of errors, and much more. At one extreme, the relationship of measurement to the manufacturing processes is crucial; at the other, a grasp of the demands of the final application is critical to successful manufacturing. This is why technology transfer is the very essence of the proposed project, so that the industrial partners can enhance their own skills in addressing the marketplace, but also so that the scientific community can benefit through enhanced technical capabilities. Technology Transfer as we call it is all about people, and one of the best ways to do it is to address a common problem as a team. The central problem we address is how most effectively to measure complex parts as they are processed on the Zeeko polishing machines. With technology-transfer in view, the project focuses on developing a challenging prototype instrument which will combine two measuring methods in a novel way. The result will be a compact measuring module which will fit into the tool-holder on the Zeeko machines. This will enable a complex part to be measured as a set of overlapping patches, using the machine's 7-axis motion-system to provide the surface-scanning. It remains to take these separate patches and mathematically stitch them together.

Publications

10 25 50