To describe fixturing as the Cinderella of manufacturing is inaccurate: it is more of an ugly sister, since very few people want to take a close look. Fixturing can dominate the cost of some manufacturing jobs, and machined fixtures are so expensive that firms daren't throw them away, so long-term storage often lengthens the bill. Bizarrely, fixtures made from kits are also often stored for long periods: the difficulty of recording a fixture design made from a kit, and the time and skilled manpower needed to assemble it, are such that manufacturers are prepared to tie up these expensive modular systems indefinitely, rather than risk being unable to fulfil a repeat order. And all these problems are compounded when several identical fixtures are called for, for instance to support the palletization that is necessary to realize the benefits of FMS and other advanced machining systems. It is therefore surprising that amount of research that has been, and is being, performed on fixturing problems is tiny compared to the huge amount of work that has led to today's CAD and robotics systems. The project to develop self-locating fixtures ran at Bath University in the mid-eighties, and was sponsored by the UK Science and Engineering Research Council and Westland Helicopters Ltd. At its conclusion, efforts were made to obtain industrial commitment to our self-locating fixtures system, either in the UK or from abroad. Regrettably, these efforts were not successful. Despite the length of time that has elapsed, the fixturing problem remains the same, few developments towards solving it have taken place, and the kit developed at Bath remains a promising lever on a manufacturing problem that currently soaks up millions. As James Dyson discovered, having a commercially sound engineering solution does not ensure immediate industrial interest. However, we believe that the sheer cost of the fixturing problem will lead to intelligent people seeking for 'a better way' from time to time. We hope they will find these pages, and may become interested in developing or otherwise taking forward the self-locating fixture.
A simple CAD system was implemented to support the design of fixtures using
this kit, by performing the computations necessary to design the towers, given
the desired positions of pins and clamps. A graphical display was provided so
that interfering towers could be detected. (Such problems can usually be
avoided because a given position can be achieved starting from a range of
baseplate positions.) However, the computer system used is now obsolete and,
although innovative at the time, most aspects of the implementation have been
superseded by facilities in commercial CAD systems and by recent research on
fixture design.
Other research on fixtures There has been a relatively large amount of research on the computer-aided design of features, although at best software can only provide a partial solution to the fixturing problem as a whole. The physical difficulties have been addressed by various conformable and multi-phase devices. These are especially useful for simple, thin-walled components, such as compressor blades, but would be difficult to integrate into a general machine shop or an FMS. Partial robotic assembly of existing fixture kits has also been investigated, but this still leaves the difficult business of dimension setting to a fitter.
|
Following a survey of parts at our industrial collaborator, the system was
designed to deal with '2½D' components, which need to sit on flat pads
of various heights, and achieve XY location by holes and edges.
The fundamental elements of the kit are towers, each consisting of three
meshing toothed discs, and a small number of special washers. The teeth on
the discs have pitches without a common factor, and thus an angular resolution
of 1/40 of a degree can be achieved by appropriate assembly. The thickness of
the washers providing height adjustment are chosen to be a solution to a
problem from number theory, called the postage stamp problem. This gives the
minimum different number of washers for a given range and resolution. Four
washers chosen from a small set can provide vertical resolution of 0.02mm.
The towers plug into a baseplate with a hexagonal grid of mounting positions:
this sort of arrangement is common to many manually assembled fixturing kits.
To provide XY locations that do not correspond with positions on the
baseplate, two towers can be connected using a short arm. Appropriate fixing
pins, clamps etc are then mounted on the second tower.
The assembly machine is a specialized pick-and-place robot of radial design,
which works in 2½D, like the fixtures themselves. It is capable of
assembling a multi-element fixture in a few minutes.