How the choice of stylus ball influences the measurement result
The same measurement with a different stylus ball returns a different value. For flatness, straightness and roundness, the numerical result depends directly on the ball diameter, and with tight tolerances it helps decide between conforming and non-conforming.
Metrology
Christian Schärer · 14 August 2026
The effect is purely geometric. A ball of radius r touches an integral feature only where the curvature allows it. Over narrow recesses it rests on both flanks and bridges them. Its centre always stays exactly r away from the feature, and that centre is what the measuring machine records. The ball therefore acts as a filter that no software setting can undo.
How the graphic works
In the graphic below, three ruby balls travel one after another across the same profile. Each draws its centre-path line in its own colour. The lines stay on screen until the run is repeated, so the three traces can be compared directly. You can set the three diameters and the actual deviation of the profile in three levels.
What the graphic shows
With a small ball, the centre-path line still follows the deviation of the feature almost everywhere. As the diameter grows, the ball bridges ever wider recesses and the recorded line becomes flatter. At 5 mm it rests only on the widely spaced peaks. None of this is a calculation error — it is the geometry of the probing element.
The ball diameter is therefore a filter parameter in its own right. It acts before any software filter setting comes into play, and it cannot be reversed afterwards: whatever the ball has bridged is no longer present in the measurement data. Metrologists call this effect a morphological filter.
Form deviations: where ball choice acts directly
In tactile measurement, choosing the right stylus ball diameter is therefore essential. Flatness, straightness and roundness are evaluated on the recorded trace, and on the same feature they come out smaller when a larger ball is used. The table below the graphic puts numbers on it. Two results are only comparable when the ball diameter is the same.
With generous tolerances the difference has no consequences. With tight ones it becomes a share of the tolerance: if a flatness tolerance sits in the range of a few hundredths and the larger ball smooths away a substantial part of the deviation, the same part can be measured as conforming with one ball and non-conforming with another. Both measurements are performed correctly, yet they are not comparable.
For straightness, ISO 1101 applies. The decisive point in every case: a form value belongs with its measurement conditions, and that includes the ball diameter. In form plots and deviation analyses the ball radius should therefore always be stated. Without it, the plot is hard to interpret, because it remains open what the ball has already filtered out.
In contour measurement the connection is even more direct. What is recorded is the centre path of the stylus arm. Radii, angles and contour segments only emerge after the stylus radius correction, and that correction can only be as good as the radius is known. At internal radii smaller than the stylus radius, no measured values exist at all, because the ball cannot reach in.
No software filter can undo it
A software filter, such as the Gaussian filter, averages the recorded trace and can be adjusted. The ball filters beforehand and purely geometrically — like a mating part in assembly, which never reaches fine recesses either. What it has bridged, no evaluation brings back.
