A microcontroller listening for trouble inside an aging production printer
A print shop mounted a small sensor board inside an old Xerox iGen4 to learn its normal vibration and flag anything unusual, plus monitor solvent fumes.
Evidence: Vendor tutorial based on a real print-shop scenario. No downtime results published. Source date: not stated · Checked 2026-10-23 · Tech: Sony Spresense microcontroller, CommonSense sensor board, Edge Impulse
The problem
An essential, aging machine with blind spots its own error screen couldn't catch, and no budget to replace it.
What they did
They recorded vibration while the machine ran normally, idled and was off, then trained an anomaly-detection model that runs on the microcontroller itself. A second model watched an air-quality sensor for ink or solvent leaks.
Reported results
The write-up is a build guide, not an ROI story: it shows the model working and is honest that the leak detector missed about 31% of leak samples in testing (few false alarms, but not very sensitive).
Our take
The key idea for shops: you don't need examples of failures to start. You train on "normal" and alert on "different." That's how you'd approach an old spindle, pump or compressor too. We like that the source shows its weak spots.
Try this on Monday: Pick one machine whose breakdowns hurt most and that has no built-in monitoring. That's the candidate for a $100-ish sensor pilot.
Brian Roemmele reports wiring an AI agent into an existing flexographic carton-box printing line, with small local models watching the machine and cameras feeding it visual data.
Evidence: Self-reported by the builder on X. Early, no independent verification yet. · via @BrianRoemmele
Brian Roemmele showed a Raspberry Pi with an attached camera running the open LLaVA 1.5 model locally to describe and recognize what passed in front of it.
Evidence: Hobbyist demo shared on X (2023). Not a factory deployment. · via @BrianRoemmele
Neo Industries, a small machine shop in Colleyville, Texas, used automatic production monitoring to see how much of each cycle every tool was responsible for, then optimized tool by tool.
Evidence: Vendor-published customer story with named shop owner.