Coupling Design Process: Improvement Techniques

After designing the coupling, several factors can be changed to improve the performance of the coupling. The basic improvement techniques are as follows:

Assembly Techniques - A variety of techniques can be used during assembly to improve performance. Repeatability has been shown to increase significantly by switching from a basic installation procedure to a refined procedure.

The basic installation procedure consists of:

1. Gross pre-alignment of coupling surfaces
2. Direct application of preload using full torque on preload bolts
3. Dry bolts and contact surfaces

The refined procedure includes:

1. Gross pre-alignment of coupling surface followed by minor adjustments of alignment to allow for optimal, non-damaging contact between surfaces.
2. Stepped application of preload (ie. 10% - 50% - 100%) in a proper, consistent pattern. Stepped application controls initial contact of the interface and reduces the initial static friction in the bolts. Torque pattern should be designed to prevent asymmetric bending across the interface and remain consistent through each bolt application.
3. Lubricated bolts and contact surfaces to reduce frictional nonrepeatability on interfaces. Grease is suggested for bolts while spray lubricants or coatings are suggested for critical interface surfaces to prevent the introduction of unknown particles.

Calibration - A variety of different calibration methods can be used to remove the errors caused by many of the disturbance actions. A basic calibration consists of measurements of the various coupling features combined with intelligent storage of correction parameters based on the measurements. Direct measurement of the contact points on the halves of a kinematic interface can greatly reduce the effect of tolerance errors on mounting accuracy, with the residual interchangeability error based only on the error of the measurement procedure. By estimating the total mounting accuracy of a kinematic coupling as the sum of the measured repeatability and the simulated interchangeability, interface manufacturing tolerances and the complexity of the calibration process can be chosen to satisfy the accuracy requirement at minimum cost. While past laboratory measurements of kinematic couplings have shown micron-level repeatability at relatively small scales, in a test application to industrial robot base and wrist mountings measured interface repeatability is approximately equal to simulated interchangeability. The interface transformation has the potential to become a universal kinematic handshake between kinematically coupled objects, and could enable a conceptually new interface-centric calibration process for modular machines, whereby:

1. Interface halves are pre-assembled and encoded with their coupling calibration information, relative to their centroidal coordinate frames.
2. These calibrated interface halves are attached to machine modules and the modules are calibrated by mounting the assembly to a reference mating interface half. The coupling parameters of the reference interface are known; hence a calibration Tinterface is known.
3. When the machine modules are brought to the production installation site, the production Tinterface is calculated from the coupling parameters of both production interfaces. A correction is applied to the machine module calibration for the difference between the calibration Tinterface and the production Tinterface. This would allow the machine to be more accurately programmed off-line.

In production, by making the contact surface measurements ahead of time, calculation of Tinterface would be a step of the machine calibration routine. Ideally, the software would take the measurement values for the components, calculate the interface HTM, and apply it to the global serial chain of transformations for the machine kinematics. The pre-measured placements of the contacts could be written to an identification tag on the interface, or the interface serial number could serve as a database key to the calibration data.

See Hart's Thesis for more information on calibration of coupling interfaces.

Iteration to a new design - If acceptable performance is not acheived using the desired coupling type, further iteration of the various design parameters should be performed. A small change in one parameter, such as a surface coating to reduce friction, may sufficiently improve performance.

Back to Coupling Design Analysis