// Showcase
Two internal faces inside an extruded aluminium profile, about 6 mm apart, to be measured to 1.667 µm at one sigma – at three positions, over lengths up to 2000 mm, and the profile must not be clamped to do it.
Two internal faces inside an extruded aluminium profile, about 6 mm apart, to be measured to 1.667 µm at one sigma – at three positions, over lengths up to 2000 mm, and the profile must not be clamped to do it. Unclamped it sags, and the sag grows with the fourth power of the length. It is warped anyway; that comes with the heat treatment. And there is exactly one place to put a sensor: the 6 mm gap itself.
Quantify every disturbance first, measure second. Temperature, split into absolute positioning (46.8 µm/K over 2000 mm), relative positioning (46.8 nm/K) and the measurand itself (140.4 nm/K). Modal analysis of the unclamped profile: first bending mode at 22 Hz – the thing rings at anything. Then four measuring principles compared; tactile dropped out for want of a usable arrangement, chromatic confocal won.
A clear, reasoned no to the first approach – phrased so you know what caused it. The sensor itself scatters below 1 µm. The measurand came out at 3.5 µm, relative sensor positioning at 2.5 µm, together 4.3 µm against a required 1.667 µm. The shortfall was not in the optics but in the coupling to the part. Plus two remarks on the drawing: the form and position tolerances carry no length reference – the longer the profile, the harder the requirement gets.
A report with a complete disturbance analysis, a modal analysis and evaluated trials. And one detail worth knowing in advance: in the split-beam arrangement the two evaluation peaks merge exactly when both distances are equal – the setup fails in precisely the perfectly centred position.
Since no sensor conforming to the installation space exists on the market, we built one: two gradient-index lenses, three spacers, focus about 100 µm, measuring range about 400 µm. The first prototype was pure noise – cause: 4 % reflection loss on uncoated lenses. With a visible-band coating, below 1 %; lens spacing at least 30 mm; a divergent beam rather than 1:1 imaging. On the optical bench the setup reached a peak-to-peak value below 0.2 µm over more than 8000 samples – about 0.1 µm at ± 3 sigma.
Get in touch
You bring the challenge. We develop the solution. From the first idea through concept and development to the validated machine, we guide your project.