Live data flowing in 7 days. No IT department required. See exactly what it looks like for your machines.
Calculate Your Downtime CostPractical guides, benchmarks, and shop floor strategies — no fluff, no spam, unsubscribe anytime.
The TPM Six Big Losses framework applied to CNC machining. Equipment failure, setup time, micro-stops, reduced speed, defects, and startup losses — with data on where most shops lose the most.
Reactive vs preventive vs predictive maintenance for CNC shops. What signals to monitor, how vibration analysis works, real-world savings for a 5-machine shop, and common pitfalls to avoid.
Published OEE benchmarks are misleading because they mix industries. Here are CNC-specific benchmarks by machine type — turning, milling, multi-axis, Swiss — with realistic percentile ranges.
Lights-out machining is the closest thing to free capacity a CNC shop can get. No operator wages. No shift premiums. Just machines cutting metal while everyone is at home. The catch: you cannot run machines unattended if you have no way to know when something goes wrong. That is where monitoring stops being optional and becomes the prerequisite.
This article covers what you need to monitor, how to build confidence incrementally, and the ROI math that makes lights-out machining one of the highest-return investments a small shop can make.
The reasons are straightforward:
Running unattended without monitoring is gambling. A broken tool at 2 AM does not just ruin one part — it can crash into the workpiece, damage the spindle, flood the shop with coolant, or run dry and start a fire. Every one of these has happened in real shops.
The question is not whether you can afford monitoring for lights-out operation. The question is whether you can afford a lights-out failure without it. A single unattended crash can cost more than a year of monitoring — $8,000 to $50,000 depending on what fails and what it damages.
Monitoring is not a feature of lights-out machining. It is a requirement.
A broken tool in attended operation means the operator hits the stop button. In unattended operation, a broken tool means the machine keeps going — now cutting air, cutting with a stub, or ramming into the workpiece. Spindle load monitoring detects tool breakage within seconds by identifying the sudden drop in cutting force. The system stops the machine and sends an alert.
Running dry is the fastest path to catastrophic damage. Coolant level sensors in the sump and flow sensors on the delivery line confirm that coolant is reaching the cut. A low-level alarm at 2 AM is an inconvenience. Running a carbide tool dry into hardened steel at 2 AM is a $15,000 problem.
Chips pack into the workholding, wrap around the tool, or pile up until they interfere with the next operation. A chip conveyor that jams during unattended operation will eventually stop the machine — but not before the damage is done. Monitoring the chip conveyor motor current detects jams before they cascade.
Vibration monitoring serves double duty during lights-out: it detects both tool wear progression and mechanical issues. A vibration spike above the ISO 10816 threshold during an unattended run triggers an immediate stop and alert. During attended operation, this might be a “watch next shift” item. During unattended operation, it is a “stop the machine now” event.
Power draw is the simplest form of machine state monitoring. Machine running = power above idle. Machine stopped = power at baseline. Machine drawing excessive power = something is wrong. A current clamp on the main disconnect gives you all three states with zero integration to the CNC controller.
Smoke detection and temperature monitoring in the machine enclosure are non-negotiable for true lights-out. A dry cutting condition with a flammable coolant mist is a fire risk. An oil mist collector that fails is a health hazard when the morning shift walks in. These are low-cost sensors with very high consequence if missing.
In attended operation, the operator IS the alert system. In unattended operation, the alert system replaces the operator. That means:
Nobody goes from fully attended to full lights-out overnight. The smart approach is incremental:
| Stage | Duration | What Runs Unattended | Prerequisites |
|---|---|---|---|
| 1. Extended Lunch | 1 hour | Proven jobs, long cycle times | Basic monitoring + alerts |
| 2. End of Shift | 2-4 hours | Stable operations, bar work | Tool breakage + coolant monitoring |
| 3. Ghost Shift | Full shift | Dedicated machines, proven programs | Full sensor suite + escalation chain |
| 4. Weekend Runs | 24-48 hours | Long production runs, bar feeders | All above + fire detection + remote camera |
Each stage builds on the last. You learn what fails, what false alarms look like, and how your machines behave when nobody is watching. By the time you reach Stage 3, you have weeks of data and confidence that the monitoring system works.
Most shops we work with reach Stage 2 within the first month. Stage 3 takes 2-3 months. Stage 4 requires the most proven processes and the most reliable monitoring, so it comes last — typically 4-6 months in.
Adding an unattended shift to even one machine is a significant capacity gain. Here is what it looks like for a single CNC machine:
| Metric | 1-Shift (Attended) | 1-Shift + Ghost Shift |
|---|---|---|
| Production hours/day | 8 hrs | 12-14 hrs |
| Production hours/year | 2,000 hrs | 3,000-3,500 hrs |
| Capacity increase | Baseline | +50-75% |
| Additional revenue (at $150/hr) | -- | $150,000-$225,000/yr |
| Additional labor cost | -- | $0 |
| Monitoring cost (per machine) | -- | $960/yr |
A ghost shift on a single machine adds $150,000-$225,000 in potential revenue per year at near-zero marginal labor cost. The monitoring that makes it possible costs $80/month per machine. That is a 150-230x return on the monitoring investment.
Even at 50% utilization during the ghost shift — accounting for cycle gaps, material changes, and conservative scheduling — the numbers are compelling. Half of $150,000 is still $75,000 in additional capacity from a single machine.
Not every machine and not every job is suited for unattended operation. The best candidates share these characteristics:
Plan your monitoring setup
Tell us about your machines, your production mix, and your lights-out goals. We will design a monitoring plan that gets you from Stage 1 to Stage 3 in 90 days.
Calculate your downtime cost →Lights-out machining is not about technology — it is about confidence. Confidence that the machine will not crash, that a broken tool will stop the program, that a coolant failure will trigger an alert before it becomes a disaster.
That confidence comes from monitoring. You build it incrementally: one hour unattended, then four, then a full shift. Each stage proves the system works, and each stage adds capacity without adding labor.
Your machines are already there. They do not need a night shift. They need you to stop turning them off at 5 PM.