Reverse engineering legacy tooling without the guesswork
When the drawings are gone and the engineer has retired, worn geometry still carries the design intent — if you know how to read it.
Scanning is measurement, not engineering
A point cloud tells you what the part is today, including fourteen years of wear. It does not tell you what it was designed to be. Treating a scan as the answer reproduces the wear along with the geometry.
The scan is an input. The engineering judgement about which surfaces are functional, which are cosmetic and which have degraded is where the value sits.
Read the mating parts
The most reliable evidence of original intent lives in the components the part touches. Bearing bores, register faces and locating pins carry the tolerances that mattered. Everything else is usually free.
Wear patterns are informative too: a polished band on a guide surface tells you where load was carried, and often that the original clearance was tighter than what remains.
Rebuild parametrically
Surface-fitting a scan produces a model nobody can modify. Rebuilding parametrically from measured critical dimensions produces a model your team can iterate on for the next decade.
It takes longer initially and pays back at the first revision.
- Establish datums from functional surfaces first
- Rebuild features in manufacturing order
- Round nominal dimensions to sensible values where wear permits
- Document every assumption in the drawing notes
Document the assumptions
Every inferred dimension should be recorded with its basis. When the first article comes back out of specification, that record is what lets you find the wrong assumption in an hour instead of a week.
Key takeaway
Scan for measurement, engineer for intent, rebuild parametrically, and document every assumption you make along the way.