🤖 LIGHTS-OUT · ROBOT CELLS · INDUSTRY 4.0

Automation & Deep Hole
Drilling Cells

Deep hole drilling automation moves parts and manages the process while the machine cuts. Cells range from a single robot loading one gundrill to fully unmanned multi-machine lines that run second and third shifts with no one on the floor. The economics hinge on recovering idle spindle time — most shops believe they run at 80% utilization when the reality is closer to 40%.

>90%UptimeRobotic cell efficiency
22–23Hrs/dayLights-out spindle time
12–18 moPaybackLevel 2 robot cell
384 mmDepthRobotic shaft cell

Five Levels, Manual to Lights-Out

Match automation complexity to part geometry, batch size, and the hours of unattended operation you actually need. Most manufacturers start at Level 2 and expand incrementally.

LevelDescriptionLoadingBest ForInvestmentSpindle hrs/dayROI Period
L0 ManualManual load, manual drill cycle — every part attendedHandPrototype, R&D$0 (existing machine)4–6N/A
L1 Automated cycleManual load, automated drilling cycleHandLow volume (<50 parts/yr)$0 (existing machine)6–7N/A
L2 Robot + palletRobot loads from pallet or grid plate, single machineRobotMedium batch (50–500), one part family$50k–$100k~12–1412–18 months
L3 Multi-machineRobot serves several machines with a conveyor systemRobot + conveyorHigh volume (500–5000), multiple operations$150k–$300k16–2018–24 months
L4 Lights-out cellFully automated cell: gantry, washer, CMM, tool presetter, auto changersFull systemHigh volume (>5000), lights-out production$300k–$800k+22–2324–36 months
💡 Perception vs. reality: Most shops believe they run near 80% utilization; the measured average is closer to 40%. Top-quartile shops run machines 14 hours/day versus 8 hours/day for the bottom quartile. Automation lifts a single machine from 4–6 spindle hours/day (manual) toward 22–23 hours/day (full lights-out FMS).
⚠️ Level is a process decision, not a budget decision: A Level 4 cell is only economical if you can keep tooling, coolant, and chips reliable for the whole unattended window. Automation amplifies a broken process — it does not fix one.

Why Automation Pays — The Utilization Math

Deep hole drilling machines are expensive, high-value assets that sit idle 16+ hours a day on a single shift. Unattended machining is the most direct way to turn that idle time into production.

Machine situationTypical utilizationSource basis
Manually attended machine30–40%Datanomix utilization studies
Shops running pallet / robot loadingapproaching 95%Pallet-system operators
Bottom-quartile Top Shops~8 hrs/dayModern Machine Shop Top Shops survey
Top-quartile Top Shops~14 hrs/dayModern Machine Shop Top Shops survey

The profit-margin ladder (Top Shops survey, 1,100+ shops)

ApproachMedian profit margin
Baseline (no improvement methodology)~1%
+ Improvement methodology~5%
+ Unattended machining~8%
+ Machine monitoring~9%
💰 Worked example: Adding 6 unattended hours per night at 70% utilization, with a $110/hour machine contribution, over 240 operating nights per year creates roughly $110,880 of annual upside — before scrap and alarm-response costs. The ROI decision lives or dies on actual uptime, not theoretical throughput.
🤝 Labor context: Chronic skilled-labor shortages make unattended shifts the cheapest new capacity available. Roughly 313,000 durable-goods manufacturing jobs sat unfilled in early 2025, and Deloitte projects nearly 2 million unfilled manufacturing jobs nationally by 2033.

Anatomy of an Automated Deep Hole Cell

A cell is more than a robot bolted beside a machine. Every subsystem must be sized and interfaced so the machine never waits and never runs blind.

Sub-systemFunction in the cellTypical hardware
Machine coreDrill the holeGundrill, BTA or ejector machine with through-spindle coolant, feed and spindle drives sized for unattended cycles
Robot / handlingMove parts between stations6-axis articulated robot (FANUC, ABB, KUKA), gantry robot for long shafts, collaborative robot for small parts
Workpiece storageBuffer raw and finished partsGrid plates, pallet systems, bulk feeders, walking beams, multi-level discharge magazines
Fixture & guide bushingLocate the part, guide the drill at entryHydraulic/mechanical clamps, entry guide bushing, bushing-changer station
Coolant & chip handlingSupply coolant, remove chipsHigh-pressure pump, filtration (10–50 μm), hinged-belt or scraper chip conveyor, tramp oil separator, level/temp monitoring
Tool managementChange, track, and preset drillsAutomatic tool changer with RFID, offline presetter, spindle-load wear monitoring
InspectionVerify the hole in or near the cellPost-process CMM or plug gage, bore scope station, coolant pressure/flow monitoring
Cell controllerCoordinate machine, robot, and peripheralsPLC/robot cell controller, OPC-UA or MTConnect connectivity, OEE data logging
10–20
μm
Gundrill coolant filtration
30–50
μm
BTA coolant filtration
25 kg
Payload
FANUC M-20iD robot, 1831 mm reach
±0.02
mm
Robot repeatability
176 pos
Grid plate
Shafts 20–35 mm (Ott-Jakob cell)
88 pos
Grid plate
Shafts 35–50 mm, up to 550 mm long

✅ Single robot, one machine

  • Lowest capital entry ($50–100k)
  • Fastest payback (12–18 months)
  • Process is proven before you scale
  • One part family — low programming risk

🤖 Multi-machine cell

  • Robot + conveyor cost spread over several machines
  • Throughput per square foot maximized
  • Parallel operations (drill + ream + finish) run simultaneously
  • Unattended shifts amortize the capital faster
  • Trade-off: higher investment, longer commissioning, and a robot outage stops the whole cell

Part Loading & Workpiece Storage Strategies

MethodBest forTypical capacityField example
Grid plateShafts, medium batch176 positions (20–35 mm) + 88 positions (35–50 mm)TBT ML250 + Halter LoadAssistant (Ott-Jakob)
Pallet system / pallet changerHigh variety, multiple part familiesMultiple pallets swapped at cycle timeUNISIG flexible pallet changers
Bulk feeder (vibratory / step)High volume, simple geometryHours of unattended operationUNISIG R-4-2-2 barrel cell
Walking beam / conveyorInline shaft productionContinuous flow between stationsPrecihole shaft gun drilling lines
Multi-level discharge magazineFinished partsAuto-stacks machined partsUNISIG R-4-2-2 finished-barrel discharge
1
Measure the part in the gripper

An integrated sensor re-checks shaft position and re-measures length before loading — drilled parts can drag out of clamping position, and an out-of-position load is a guaranteed crash.

2
Keep the machine cutting

Stage parts in a buffer station in front of the machine so the robot can load four at a time while the drill runs. The TBT/Ott-Jakob cell loads four shafts per cycle to cut machine idle time to near zero.

3
Account for wet weight

Robot payload must include the weight of coolant clinging to a freshly drilled part. A part that reads dry and leaves dripping can stall the gripper at full extension.

4
Handle the guide bushing

In gundrill cells the entry bushing defines hole position. When diameter changes, the cell must swap bushings or re-machine the seating — a bushing-changer station keeps the cell running between part families.

5
Design for expansion

Overspecify the cell controller and coolant system capacity from day one. Adding machines, inspection stations, and cleaning stations later is cheap only if the backbone was built for them.

💡 Start small, scale up: Most manufacturers run a single robot serving one deep hole machine first, then expand to a multi-machine cell once the automation is proven. A modular approach lets you add machines, inspection, and cleaning incrementally.

Gundrill vs. BTA vs. Ejector in an Automated Cell

The drilling method shapes the cell as much as the robot does — coolant pressure, chip volume, and tool-change frequency drive the automation design.

Design factorGundrill cellsBTA (STS) cellsEjector (DTS) cells
Diameter range0.5–50 mm6–2000 mm18–250 mm
Coolant pressure50–150+ bar15–100 bar10–50 bar
Filtration target10–20 μm30–50 μm≤30 μm (protect Venturi)
Tool change frequencyHigh — regrind after short depthLower — indexable insertsModerate
FixturingEntry guide bushing mandatoryPressure head seal at workpiece faceNo face seal — irregular faces OK
Chip loadFine, continuous chipsLarge-volume C-chipsC-chips; Venturi needs clean coolant
Robot / loading notesLight parts, frequent changeover, bushing handlingHeavy bar loaders, swivel unitsLowest entry cost; retrofits onto existing machines
Automation complexityHighest (tooling + coolant + bushing)MediumLowest entry cost
⚠️ Guide bushing handling is the hidden chore: A gundrill's entry bushing sets both position and coolant seal. In an unattended cell, diameter changes mean bushing swaps — budget for a bushing-changer station or a robotic bushing-seat machining operation, or the cell stops dead at the first family change.
💡 Method → cell decision: Gundrill cells pay for tooling management and bushing automation. BTA cells pay for heavy loaders and big chip conveyors. Ejector cells are the lowest-cost automation retrofits because they drop onto existing lathes and machining centers with no pressure head.

Monitoring & Machine Intelligence

Unattended operation is impossible without sensors watching the process. Coolant and spindle signals are the eyes of a deep hole cell.

Monitored signalHow it is measuredWhat it catches
Coolant pressurePressure transducer on the supply lineSudden drop = chip blockage or pump failure; ejector Venturi loses chip evacuation below ~8 bar
Coolant flowFlow meter (e.g. Guhring PQ 3000) with tool clamped in spindleClogged filters, pump/seal wear, narrowed coolant ducts, drill breakage
Spindle load / thrust powerDrive feedback or load cellTool wear, tool breakage, chip jamming
VibrationAccelerometer on spindle housingAbnormal gundrill vibration before breakage — critical for small-diameter gundrills (≤4 mm)
Axial forceLoad cell in the feed axisChip evacuation failures during the cut
💡 Close the loop: DMG MORI's adaptive drilling control (ADC) continuously reads spindle load, coolant pressure, and flow and adjusts parameters in real time — reported roughly 30% longer tool life and comparable energy savings versus rigid-parameter cycles. The same signals feed OEE and predictive maintenance via OPC-UA or MTConnect.
🔥 Flow beats pressure: Next-generation gundrilling machines (e.g. the UNISIG UNE series) switch from pressure-based to programmable flow-based coolant delivery, which operators say lets them predict tool breakage and cut downtime from interrupted processes. Plan the cell controller to capture these signals, not just the cycle-complete flag.
30%
Tool life
Adaptive control gain (DMG MORI ADC)
≤4 mm
Gundrill
Vibration monitoring is mandatory below
~8 bar
Min pressure
Ejector Venturi evacuation floor
OEE
Metric
Logged per machine from MTConnect data
29%
Downtime cut
Via predictive spindle-load alerts (sfHawk case)
18%
Tool cost
Reduction via optimized tool-life monitoring

Three Production Cells in the Field

Example 1:Robot-assisted gundrill (TBT ML250 + Halter LoadAssistant)
Site:Ott-Jakob Spanntechnik, Lengenwang, Germany — spindle coolant holes for tool-clamping systems
Robot:Halter LoadAssistant Universal Premium 35, double-sided end effector with workpiece-specific gripper fingers
Storage:176-position grid plate (20–35 mm) + 88-position grid plate (35–50 mm), shafts 75–550 mm long
Loading:Four parts at a time through a buffer station in front of the machine; sensors re-check position and shaft length to prevent collisions
Result:In service since March 2023, ~3.5-minute loading cycle, a full 8-hour shift plus an unattended second shift
Example 2:Multi-machine barrel cell (UNISIG R-4-2-2)
Machines:UNI four-spindle gundrill + UNR reaming machine + R-series button-rifling machine
Handling:Smart conveyor draws barrel blanks from a bulk feeder; 6-axis robot transfers parts between drilling, reaming, blow-off, and rifling; finished barrels to a multi-level discharge magazine
Operation:Four barrels gundrilled while two are precision reamed and two receive button rifling — all operations run simultaneously
Result:Hours of unattended operation; multi-spindle design multiplies productivity without adding floor space. Previously affordable only by the largest firearm OEMs.
Example 3:Automotive transmission shaft cell (Mollart + FANUC)
Machines:Three-machine production cell for hydraulic gearbox transmission shafts
Robot:FANUC M-20iD/25 6-axis robot on a gantry rail — 25 kg payload, 1831 mm reach, ±0.02 mm repeatability
Control:FANUC Series 35i-Model B (up to 20 axes), Industry 4.0 networking with OPC-UA, machine light stacks, ERP integration, per-machine energy monitoring
Drilling:Deep holes to 384 mm depth at 200 mm/min, within 0.2 mm at the start and 0.3 mm deviation at the end
Result:>90% uptime efficiency; unattended second shift is standard; single warranty across robots, controls, and networking via one provider

Building the Business Case

Calculate ROI on actual utilization, not theoretical maximum throughput. Include programming time, maintenance, and the learning curve for setup personnel.

1
Measure current utilization

Spindle-on time versus clock time over a real production week. Most shops overestimate — the perceived 80% is usually closer to 40%.

2
Define the target scenario

How many unattended hours per night or shift, and at what planned utilization. Proven processes justify higher targets.

3
Cost the cell honestly

Hardware + integration + programming + training + ongoing maintenance and consumables, not just the robot price.

4
Model the margin

Contribution per machine hour × added productive hours × operating nights per year.

5
Add the risk factors

Scrap events, alarm-response time, commissioning duration, and the setup learning curve all eat into the model.

6
Compare to labor

Second-shift operator cost versus robot total cost of ownership. A fully loaded CNC operator runs $55k–$75k/year.

Machine contribution:$110/hour
Added unattended:6 hours/night at 70% utilization
Operating nights:240 per year
Annual upside:~$110,880 before scrap and alarm-response costs
🤖 Cobot economics: A robotic cell that cost $200k in 2020 now runs $80k–$120k, while a loaded CNC operator costs $55k–$75k/year. Payback crossover is now as low as 18 months for single-shift and 8–12 months for two-to-three-shift operations.
⚠️ When NOT to automate: For batch sizes under 50 parts, setups changed more than once per shift, or extremely long cycle times over 30 minutes per part, manual loading is often more economical. Unattended machining adds no profit margin without a foundational improvement methodology underneath it.

The Unattended-Shift Checklist

Run these conditions before you trust the machine to work alone. Each one is a failure mode that scrapes an entire night of production if it breaks.

1
Prove tool life

The drill must survive the full unattended window. Monitor spindle load and program automatic retract on threshold before breakage.

2
Guarantee chip evacuation

Chip-conveyor capacity and coolant filtration must handle peak chip flow. A jam four hours into a lights-out run is a full scrap event.

3
Control coolant

Automatic level make-up, temperature control, tramp-oil separation, and pressure/flow alarms with remote notification.

4
Verify the first part

Probe hole position and depth on the first part of every shift; plug-gage or bore-scope critical features before letting the cell continue.

5
Plan alarm response

Remote alarming to a phone, written recovery procedures, and clear instructions for the morning shift about what to inspect first.

6
Install fire suppression

High-pressure oil mist is a fire risk. CO₂/inert-gas suppression, mist extraction, spark detection, and an overpressure-rated enclosure.

✅ Validate one process at a time: Successful shops start with a single proven part family, run it unattended, and scale from there — lights-out eventually accounts for 30–50% of total production hours.

Key Safety Points for Automated Cells

🔥 Oil mist fire risk: High-pressure cutting oil creates an explosive mist inside the machine enclosure. Every automated deep hole cell needs mist extraction rated for explosive atmospheres, spark detection with automatic suppression, and an overpressure-rated enclosure. A 1 mm oil film inside the enclosure burns at 800°C — clean mist accumulations on a fixed schedule.
⚠️ High-pressure coolant: 50–150 bar lines are lethal if disconnected under pressure. Relieve at the pump before maintenance, use whip-checks on every high-pressure hose, and never defeat interlocks.
⚠️ Robot cell guarding: Interlocked perimeter fencing or light curtains around the robot envelope, automatic doors that only open in the safe zone, and a single E-stop chain covering robot + machine + conveyor.
⚠️ Chip conveyor pinch points: Lock out and tag out before clearing conveyor jams — swarf is razor-sharp and conveyor chains do not wait for operators.

Keep Reading