.png){kind=logo}
When manufacturers decide to invest in machine monitoring, one of the first practical questions they face isn't about software features or dashboards — it's about connectivity. How does the monitoring system actually get data from the machine?
There are two primary approaches: hardwired (direct) connectivity, which integrates directly with the machine's controller, and sensor-based connectivity, which uses external devices to infer machine state without touching the controller. Both methods work. Both have real trade-offs. And for many shops, the right answer is a combination of both.
This article breaks down how each method works, what data each can provide, and how to decide which approach makes the most sense for your equipment and your goals.
Hardwired connectivity means the monitoring system communicates directly with the machine's CNC controller, PLC (Programmable Logic Controller), or other embedded control system. Data flows from the machine's brain to the monitoring platform through a physical or network-based connection.
Common hardwired integration methods include:
• MTConnect: An open, royalty-free standard developed specifically for CNC machine tool communication. MTConnect-enabled machines broadcast a structured data stream that monitoring systems can read directly. Most modern CNC controllers from Fanuc, Mazak, Okuma, Haas, and others support MTConnect natively or via an adapter.
• OPC-UA (Open Platform Communications Unified Architecture): A machine-to-machine communication protocol widely used in industrial automation. OPC-UA is more flexible than MTConnect and is common in PLC-controlled equipment, robotics, and broader factory automation environments.
• FOCAS / FANUC Open CNC API: A proprietary Fanuc interface that allows direct communication with Fanuc controllers, commonly used in shops with large Fanuc-based CNC fleets.
• Ethernet / Serial connections: Direct network or RS-232/RS-485 connections to legacy controllers that don't support modern protocols but still expose data through older communication interfaces.
Hardwired connectivity reads data that the machine itself already knows — spindle speed, feed rate, program number, tool in use, alarm codes, part counts, and cycle state. This is first-party data, coming directly from the source.
Sensor-based connectivity takes a non-invasive approach: instead of talking to the machine's controller, external sensors attach to the machine and measure physical signals to infer what the machine is doing.
The most common sensor types used in machine monitoring include:
• Current / Power sensors (CT clamps): Clip-on current transformers measure the electrical current draw of a machine's power circuit. When a machine is running under load, current draw is high; when idle or off, it drops. This allows the system to detect running, idle, and off states based on power consumption patterns.
• Vibration sensors (accelerometers): Measure mechanical vibration generated by the machine during operation. Useful for detecting whether a spindle is rotating, whether a press is cycling, or whether a conveyor is moving.
• Acoustic sensors / microphones: Capture sound signatures to detect machine activity. Less common for general monitoring but used in specific predictive maintenance applications.
• Proximity / reed switches: Simple contact or non-contact switches that detect mechanical motion — for example, detecting each stroke of a press or each rotation of a spindle to count cycles.
• Vision-based sensors: Camera systems that observe machine or operator activity to infer production state. Emerging technology with growing adoption in specific environments.
This is the most important comparison when evaluating connectivity methods. The type of connection determines what data you can collect — and data quality directly determines what decisions you can make.
The honest answer for most manufacturers is: it depends on your machines, your goals, and where you are in your Industry 4.0 journey. Here's a practical framework to help you decide.
Choose Hardwired / Direct Connectivity When...
• You have modern CNC machines (typically 2005 or newer) with MTConnect, OPC-UA, or Fanuc FOCAS support
• You need accurate cycle times and part counts for OEE, job costing, or scheduling
• You want controller-level data: alarm codes, program numbers, tool data, spindle load
• You're building a monitoring program intended to scale into predictive maintenance or advanced analytics
• Data accuracy is business-critical — for customer reporting, quality records, or quoting accuracy
Choose Sensor-Based Connectivity When...
• You have legacy machines with no modern communication protocol — older mills, lathes, presses, or specialty equipment
• You need to get monitoring up quickly across a large, mixed fleet with minimal setup time
• Your primary goal is basic utilization visibility: is this machine running or not?
• Controller access is restricted due to warranty concerns, vendor agreements, or operator resistance
• You want to add energy/power monitoring alongside basic utilization tracking
Use a Hybrid Approach When...
• Your shop has a mix of modern CNC equipment and older legacy machines
• You want high-fidelity data from your most critical or highest-value machines, and basic utilization data from the rest
• You're phasing in monitoring over time and want to start with sensors on everything, then transition key machines to direct integration as you mature
💡 The most common real-world scenario: a shop with 20 machines where 12 are modern CNC centers (candidates for hardwired integration) and 8 are older manual or semi-automated machines (better suited for sensor-based monitoring). A hybrid approach gives you rich data where it matters most and baseline visibility everywhere else.
One framework worth keeping in mind when choosing a connectivity method is data maturity. Most manufacturers move through predictable stages as they adopt machine monitoring:
• Stage 1 — Visibility: Are my machines running? How much of the scheduled time are they actually producing? (Sensor-based monitoring can answer this.)
• Stage 2 — Measurement: What are my actual cycle times? What is my true OEE? Where exactly am I losing time? (Direct integration is needed to answer this reliably.)
• Stage 3 — Analysis: Why are my machines performing the way they are? What patterns predict failures or quality escapes? (Requires rich, controller-level data over time.)
• Stage 4 — Optimization: How do I systematically improve performance, reduce variability, and integrate machine data with scheduling, quoting, and ERP? (Requires deep, accurate data across all systems.)
Sensor-based monitoring is a viable entry point for Stage 1. But shops that want to progress to Stages 2, 3, and 4 will eventually need direct integration on their critical machines. Choosing a monitoring platform that supports both methods — and allows you to upgrade connectivity as your needs mature — is an important consideration.
It's worth clarifying a common source of confusion: hardwired/direct connectivity refers to the data protocol (how data is extracted from the machine), not necessarily the physical cable running to a monitoring device. Within both hardwired and sensor-based approaches, the monitoring hardware itself can transmit data to the cloud via:
• Wired Ethernet: Most reliable for production environments; no interference, consistent bandwidth, easy IT management.
• Wi-Fi: Flexible and often sufficient for monitoring applications; requires stable shop floor wireless coverage.
• Cellular (4G/LTE): Useful for remote machines, outdoor equipment, or locations without network infrastructure.
For most shop floor deployments, wired Ethernet to a local gateway device — which then transmits data to the cloud — provides the best combination of reliability and simplicity.
• Supports both direct (hardwired) integration and sensor-based connectivity — giving you the right approach for every machine in your shop
• Connects directly to CNC controllers via MTConnect, OPC-UA, Fanuc FOCAS, and other common protocols for high-fidelity machine data
• Deploys current and vibration sensors on legacy or non-networked machines to capture utilization data without controller access
• Enables a hybrid deployment model — mixing direct integration and sensor-based monitoring across the same shop floor dashboard
• Automatically calculates accurate OEE using controller-sourced cycle time, part count, and alarm data where direct integration is in place
• Delivers consistent dashboards and reporting regardless of the underlying connectivity method for each machine
• Supports a phased implementation approach — start with sensors for immediate visibility, then upgrade to direct integration on key machines as your program matures
• Provides hands-on implementation support from a team that understands manufacturing equipment and integration challenges
Starting at $100/month per machine, Caddis makes it practical to monitor your entire floor — from your newest machining center to your oldest manual lathe — without waiting for a full fleet upgrade or a lengthy enterprise implementation.
There is no universally right answer between hardwired and sensor-based connectivity — only the right answer for your machines, your data needs, and your timeline. What matters most is choosing a monitoring platform flexible enough to support both approaches, so you're never forced to compromise on data quality for modern equipment or leave legacy machines out of your monitoring program entirely. Caddis Systems is built to meet your shop where it is today and grow with you as your needs evolve.
Want to talk through the right connectivity approach for your specific equipment? Schedule a free consultation with the Caddis Systems team at caddissystems.com.
.png)
See how Caddis can provide real-time machine insights and proven playbooks to improve your plant operations on Day 1.
Book a Demo
.png){kind=small}