Accuracy Is a Chain

A precise robot does not automatically produce a precise log component. Neither does a precise scanner.

The finished result depends on the complete manufacturing chain:

• scanning the natural surface
• scanner calibration
• processing the measured geometry
• coordinate transformations
• positioning and supporting the log
• robot calibration
• tool geometry and calibration
• machining
• movement or deformation of the timber
• measurement of the finished component

Every step can contribute to the difference between the calculated geometry and the piece of wood that comes out of production.

Three Numbers That Are Not the Same Thing

Scanner accuracy, robot repeatability and finished-part accuracy describe different things.

Scanner accuracy describes how accurately the measuring system determines the geometry of the physical log under defined conditions.

Robot repeatability describes how consistently the robot can return to a commanded position under defined conditions.

Finished-part accuracy describes how closely the manufactured timber matches the geometry we intended to produce.

Quoting the smallest of those three numbers as "our accuracy" would be convenient. It would also be meaningless.

Wood Is Part of the System

We are machining a large natural material, not a block of stable aluminium on a machining centre.

A log can move while it is supported. Its own weight can matter. Moisture can change. Internal stresses can be released when material is removed. The surface itself is irregular.

Those are not excuses for inaccurate manufacturing. They are variables that belong in any useful discussion of accuracy.

Measure the Finished Result

The number we ultimately care about is not the prettiest specification in a scanner brochure or a robot datasheet.

It is what the complete system produces on real timber.

That means comparing manufactured features with their intended geometry using a defined measurement method.

If we want to know how accurately two logs fit together, eventually we have to measure the things that determine how those two logs fit together.

Publish the Setup With the Number

When we publish manufacturing-accuracy figures, the measurement method belongs beside the result.

• What feature was measured?
• How was it measured?
• How many components were measured?
• What log sizes and geometries were included?
• What tools and machining operations were used?
• Under what conditions were the measurements made?
• Are we reporting maximum error, average error, distribution or something else?

Then we can talk about manufacturing accuracy without hiding behind a machine specification.