Three Hardware Faults That Look Like Operator Error

Roughness Testers 5 min read

Results that suddenly drift, repeat measurements that will not agree, an odd noise from the drive unit - three symptoms that usually point at the instrument, not at how you use it.

1. Sensor failure: readings swing

The sensor does the work, so check it first. Look at the probe for impact marks and at the stylus for bending. Then measure a certified reference specimen and compare against its nominal value: if the deviation exceeds the instrument tolerance, the sensor has degraded and normally has to be replaced.

There is no adjustment that recovers a worn stylus. Continued use only produces numbers that look plausible.

2. Drive unit: the trace wanders

The drive unit carries the sensor across the surface at constant speed. Worn gears or a slack belt make it stutter or jump, and what you record is no longer the real profile – it is the profile plus the mechanism.

The check takes a minute: run the unit unloaded, listen for smoothness and watch whether the probe travels at even speed. Any hesitation in the drive appears in the data as roughness that is not there.

3. Ageing electronics: zero drift

Instruments that have been in service for years drift. Cold solder joints, bulging capacitors and oxidised connectors all distort the signal between sensor and processor. The signature is unmistakable: the reading does not return to zero with no workpiece present, or repeated measurements at one position keep sliding in the same direction.

This one needs proper service. Opening the housing yourself voids the calibration and rarely fixes the cause.

The order to check in

Verify the whole instrument against a reference specimen first, then examine drive and probe separately, and only then suspect the electronics. Most "the instrument is broken" cases are resolved at step one – either the reference check passes and the problem is in the workpiece or the technique, or it fails and the sensor is the answer.

What counts as normal

Three consecutive measurements at one point, with scatter inside the instrument tolerance and a small offset from the reference value, is normal behaviour. Only a deviation that reproduces reliably, or that grows steadily with use, justifies working down the three steps above.

Related products

Need a model recommendation?

Tell us the workpiece, thickness range and accuracy you need — we will reply within 8 hours.

Ask an Engineer
Scroll to Top