// Showcase

Sources of error in placement accuracy across a square meter

We sign off a step-and-repeat machine on axis accuracy - ±5 µm over the full stroke, interferometric per ISO 230-2, passed clean.

Sensors & Metrology

Sources of error in placement accuracy across a square meter

Sensors & Metrology

01Starting point

We sign off a step-and-repeat machine on axis accuracy - ±5 µm over the full stroke, interferometric per ISO 230-2, passed clean. But the customer isn't buying an axis position - they're buying a structure that's consistent across a full square meter. In between sit two more flexible parts: the tooling foil and the substrate. If the placement error across the whole substrate turns out bigger than the axis error, there are three possible culprits - and the machine itself can't tell them apart.

02Approach

Blaming the machine would've been the obvious move - and it would've been demonstrably wrong here. Instead, we separated the causes with measurements. Back in the analysis phase, we'd already decided to pull the tooling process into the project scope, and we built the machine to measure fiducials on the tool and the substrate independently. Two reference lines across the field - one in X, one in Y - separate systematic scale error from pure scatter. We measured across several field sizes, with and without unloading and reloading.

03Result

An error budget that calls out its culprits by name - instead of one single deviation number. Tool stretching: about 10 µm in X, about 30 µm in Y - the dominant term against a ±5 µm machine spec. Tool shear angle up to 0.0105 rad. Substrate rotation 0.0645 rad, correctable via vision and the θ-axis. Substrate shear 0.0210 rad - whose cause couldn't be isolated in this campaign, and we say so straight out. From there, the effort shifted from machine calibration to the upstream tooling process.

What's inside

A scripted, reproducible analysis - raw data stays put, and charts can be regenerated from it any time, even two years later with a changed assumption. The method scales from three imprints up to a hundred, across field patterns from 3×3 to 14×4, with evaluation per measurement point, per position, and as a spatial map. Repeat and reload measurements cleanly separate the handling contribution from the process contribution.

And what happened next

The finding shifted the focus from machine calibration to the tooling process, which is why it was pulled into the project as early as the analysis phase. The vision system behind the fiducial measurement and the machine itself have showcases of their own: «Fiducial detection on mirror, glass and hologram» and «Large-area microstructure replication with micrometre precision».

All showcases

Get in touch

We are ready for new challenges

You bring the challenge. We develop the solution. From the first idea through concept and development to the validated machine, we guide your project.