A customer of ours measures and classifies a number of shims of various thicknesses and geometries during his assembly process. He currently uses two pneumatically driven probes and custom fixtures. The reaction of the probes is quite slow, the mechanical fixture is inflexible and quite expensive as it has to be customised for each shim type. Repeatability and accuracy are more than adequate with the current solution. The inductive probes in use have 0.0002 mm resolution. In the past some of the probes have been damaged by the operators - a design review should avoid all possibility of damaging the probes as one hour of disturbance in the line can far exceed the cost of the complete fixture.
The laser tested for this feasibility study was a 2D laser from micro-epsilon. The laser has a measurement range of 20 mm in Z and 20 mm in X with a resolution of 0.002 mm. The laser and its interface are fast. We could collect, via the well documented ethernet interface, a complete matrix of X/Z values at 300 Hz. This is more than adequate to locate the part and check its resting position. We chose to position the laser above the part and rest it on a narrow, slightly alleviated ridge. The ridge, a cylinder dowel, was high enough to lift the part off the surface of the fixture even if the shim was slightly bevelled. The shim, resting on this slightly raised edge, was not lying exactly parallel to the surface of the fixture. We expected to detect this position and the resting surface and correct the thickness measurement accordingly.
Minimal influence from changes in surface quality
We expected some negative influence from variations in surface quality (grinding patterns, stains and discolouration) however these were minimal as the surfaces did not range from "mirror quality" to dull or blunt.
Laser scan of the dowel
We took reference off the dowel and the base of the fixture (position in X using the dowel and height in Z using the base of the fixture). We could have also used a setting master. The shims position could be easily detected by the laser which was positioned perpendicular to the dowel. If the shim was bent, the measured scan had a gradient, however the distance between the dowel and the shim was recorded and compensated for the gradient. Gradients above a predefined limit could be rejected by a plausibility test.
The repeatability of the setup was tested measuring the same shim 25 times and logging all raw and calculated values. We used a linear regression algorithm to calculate the exact position of the top surface of the shim. We deducted the height of the dowel taking into account of its position. The maximum and minimum deviations recorded of the thickness of the shim were +0.002 mm and -0.001 mm from the mean value - a respectable result considering the enormous increase in flexibility and speed. The mean value differed only 0.002 mm from a value measured with a micrometer.
Speed and flexibility of the Laser solution
Speed is probably the greatest improvement as the fixture measures continuously. A solution has however to be found to make sure that the measurement is recorded when the operator has let go of the part. A pneumatic probe has a measurement force of approximately 100 grams. A laser has no such force and any contact between the operator and the part will have an influence upon the measurement.
The 20 mm x 20 mm window had enough measurement range to measure all shims types provided.