Phase3D's Fringe Inspection outperforms calipers on hard-to-measure AM part holes

Phase3D has published a case study on Fringe Inspection, its in-process dimensional inspection system for metal additive manufacturing, showing the tool can match manual caliper measurements with R² values above 0.99. Phase3D says the technology can "reduce reliance on manual inspection workflows, shorten qualification timelines, and lower inspection costs" on metal AM production lines.

Fringe Inspection uses structured light scanning to capture high-resolution heightmaps of both powder and melted layers as a part builds, rather than waiting for the print to finish and cool. The system can be retrofit onto existing metal AM machines or ordered on new-generation platforms. A companion software package, Fringe Operator, interrogates that heightmap data and pulls out quantitative dimensional measurements, which Phase3D compared against manual calipers and nominal CAD dimensions in the new study.

The comparison followed a Gage R&R-style protocol: three technicians each measured holes, pins, and flat wall density cubes five times apiece, using both calipers and Fringe Operator. Correlation between the two methods came in at R² = 0.9989, 0.9990, and 0.9959 for the three technicians. Engineers who inspect internal features will care most about the hole results, since holes are traditionally hard to measure with calipers and often require CT scanning to verify.

Hole features showed the largest spread between Fringe Inspection and caliper readings, which Phase3D attributes to how sensitive manual caliper positioning is on internal geometry. Even so, Fringe Operator posted lower mean absolute error against CAD than calipers did across every hole size tested. Pin measurements told a more mixed story: Fringe Operator beat calipers on medium and large pins but showed higher error on the smallest pin size, where manual edge selection becomes the limiting factor. Wall features, the simplest geometry in the test, produced nearly identical error between the two methods.

Phase3D says future development will focus on automated edge detection and feature extraction algorithms to cut operator influence and improve accuracy on small features, an area where the current caliper comparison still holds an edge. The company frames the results as evidence that dimensional validation can start moving from a post-build, manual step to something checked layer by layer while the part is still printing.