Autodesk Technology Centers Use Tulip for Machine Monitoring on CNC Equipment

Autodesk Technology Centers in Boston operate shared fabrication facilities — including woodshops, laser cutting equipment, and CNC mills — used by a wide range of organizations and operators with varying skill levels. Without real-time visibility into machine utilization, cycle behavior, or equipment health, supervisors like Workshop Supervisor Josh Aigen had no automated way to detect anomalies, track usage patterns, or alert staff when equipment needed attention. Many machines were 'stranded' — not networked and unable to report operational data. This gap made it difficult to identify bottlenecks, support new users on complex equipment, and demonstrate operational transparency to technology partners.

The Autodesk Technology Centers partnered with Tulip to deploy the Edge IO — Tulip's next-generation edge device — directly connected to a non-networked CNC mill. Off-the-shelf current clamps and vibration sensors were attached to the machine's chassis and power supply cabinet, feeding analog signal data into the Edge IO. The device runs Node-RED onboard, a flow-based programming environment that processes sensor streams locally before forwarding structured data to Tulip's cloud platform. Tulip-supplied Node-RED flows simplified sensor configuration, eliminating custom development. Processed data was visualized in frequency and time-series graphs within no-code Tulip apps, and machine learning models were trained on the captured sensor data to classify machine states and detect operational anomalies in real time.

The deployment gave the Autodesk team continuous visibility into CNC mill behavior without requiring network access to the machine itself. Key outcomes included:

• Anomaly detection via trained ML models that alert operators visually when the machine enters an unexpected state — particularly valuable for tool wear and cutting irregularities
• Real-time equipment alerts triggered by sensor events (e.g., dust collector fill level), reducing maintenance response lag
• Improved new-user support: operators with less experience receive live machine-state feedback, flattening the learning curve on complex CNC equipment
• Operational data became accessible remotely via Tulip's cloud, enabling oversight from anywhere

• Off-the-shelf current clamps and vibration sensors can unlock machine monitoring on stranded, non-networked equipment without hardware retrofits or OEM integrations.
• Edge devices with onboard processing (Node-RED) reduce cloud bandwidth requirements and enable local logic before data reaches the platform.
• ML anomaly detection is viable even with simple analog signals — the value is in pattern classification, not sensor sophistication.
• Low-code platforms allow operations teams to build and iterate on monitoring apps without dedicated software engineers or IT dependencies.
• Machine learning models trained on real operational data create feedback mechanisms that actively support operator training, not just maintenance alerting.