Rowan University and Phase3D Partner on Metal 3D Printing Quality Monitoring

Rowan University's Digital Engineering Hub has a new metal 3D printer. What makes the installation unusual is the layer-by-layer monitoring system watching it run.

The partner is Phase3D, a specialist in in-situ quality assurance for additive manufacturing. Phase3D's Fringe Inspection system uses structured light to produce calibrated heightmaps of each powder bed layer as a build progresses. The result is a live feed of the part's actual geometry against the intended geometry, not a post-build scan that comes too late to intervene.

The Day-One Test

The system earned its keep almost immediately. During Rowan's first independent build after installing a new DMG MORI LASERTEC 30 SLM US system, a preparation error caused the printer to print only part contours for the first 60 layers. The infill and support structures never loaded. That error would normally leave a team facing a ruined build, a damaged recoater, or both: transitioning from a contours-only path straight to full infill without the intended supporting geometry can cause peeling, deformation, or powder bed disruption that cascades through the rest of the print.

Instead of cancelling or crossing their fingers, the Rowan and Phase3D teams used the Fringe Inspection heightmaps to set a clear stop condition. They agreed they would abort if the maps showed part protrusions, abnormal swelling, or disrupted powder layers. When the critical transition arrived, the heightmaps showed the parts forming correctly: no swelling, no protrusion, the recoater running cleanly. The build completed successfully.

Niall O'Dowd, Phase3D founder and CEO, said the moment was exactly what the system is designed for: Day one is when inspection must prove its value. This build did not follow the plan, but Rowan did not have to choose between cancelling reflexively and hoping for the best.

Why Heightmaps Change the Calculation

Metal AM has always suffered from a visibility problem. Once the build chamber is closed and the laser is running, operators have limited information about what is happening on the powder bed. Errors that appear early -- bad supports, incorrect parameters, recoater misalignment -- can propagate through hundreds of layers before a post-process CT scan or section cut reveals them.

Fringe Inspection changes that by measuring the bed after every layer. The data feeds a go/no-go decision at the moment it matters, not after the part has already failed. For a research environment like Rowan's DEHub, where students and faculty are qualifying new processes and training the next generation of AM engineers, that real-time feedback loop is significant. It shortens the feedback cycle from days to minutes.

What It Means for Training

The partnership is not just about catching bad builds. Rowan sees the monitoring system as an educational tool. Students working with the LASERTEC system now have direct access to process data that used to require specialist equipment and external labs. Calibrated, unit-based heightmaps give them a quantitative view of how layer geometry responds to parameter changes, something that builds intuition faster than black-box results.

For Phase3D, the Rowan installation is a reference deployment in a research setting. The company's commercial work has largely targeted production environments where the cost of a failed build is measured in thousands of dollars and missed delivery dates. A university lab is a different kind of showcase: it signals that in-situ monitoring is becoming table stakes for metal AM infrastructure, not a premium add-on for the best-funded production floors.

The Bigger Picture

Metal additive manufacturing is moving from prototyping toward production at a pace that depends heavily on process confidence. Customers who want to put metal AM parts into aerospace, defense, or medical applications need repeatable, qualified processes with audit trails. In-situ monitoring systems are part of that qualification story: they provide the data required to demonstrate that a process was in control, layer by layer, throughout the build.

Rowan and Phase3D have shown one version of how that data gets used. A build went off-script on the first try, a monitoring system caught the deviation, and a team with the right information made the right call. That is a small story in isolation, but it is exactly the kind of moment that production metal AM needs to become routine.