# Two weeks down, and every sensor said fine.

Machine class (retold): Precision laser micromachining

An intermittent placement issue in one corner of the working field, escaping inspection, on a machine that passed every check it had.

- Open before Cala: Two weeks
- Hypothesis to confirmation: First check
- Signals showing abnormal: None

## The issue

A precision laser drilling machine started placing features slightly off position, but only in one region of the working field, and only occasionally. Most parts measured clean, so the escapes reached the next process step before anyone saw a pattern, and finished assemblies were scrapped downstream.

The line pulled the machine immediately. This is the kind of issue with almost no clue attached: intermittent, confined to one region, and invisible to the machine itself.

## Two weeks of clean checks

Every subsystem was checked, and every subsystem passed. Optics within specification. Motion within specification. Power, environment and alignment all normal. No error, no alarm, no trend that moved when the issue did. After two weeks the machine was, by every measurement it could make of itself, perfect. It still could not be released.

## What Cala surfaced

Cala got the symptom the way you would tell a colleague: where it happens, how often, and the two weeks of clean checks. Its top-ranked hypothesis pointed somewhere nobody had looked. Not at a component, but at a software compensation term that adjusts placement across the working field. If that term carried the wrong magnitude, the machine would behave exactly like this: healthy everywhere, wrong in one region, intermittently, with every sensor content, because it was doing precisely what it had been configured to do.

References (retold terms): Field compensation reference, rev D · p. 31; Machine configuration guide · p. 12

## The check

Read the configured value and compare it with the commissioning record. Minutes of work. No teardown, no parts.

## The outcome

The term was roughly ten times its intended value. Restored, the issue disappeared and the machine requalified. The first hypothesis on the list was the one, reached from first principles on a machine where nothing measured abnormal, after two weeks of everything else.

---

A real investigation from early testing on production equipment, retold on a different class of machine so the site, the tool and the people cannot be identified. The reasoning is the part that is true.

Canonical page: https://calalayer.com/case-studies/two-weeks-down-every-sensor-fine

Cala (calalayer.com): AI root-cause investigator for hardware engineers.