An automated screwdriving line is often described at the proposal stage as "unmanned". After ramp-up, most lines still need one or two operators. The real difference is not whether people are present, but how often they intervene, how long each intervention takes, and whether the line has to stop while they do it. Getting those three numbers down on paper is more useful than arguing over whether a machine counts as fully automatic.
1. Reduced labour and unmanned operation are different goals
The two terms are often used interchangeably, yet they imply quite different investments.
- Unmanned operation: feeding, fastening, inspection and data traceability form a closed loop; exceptions are handled or isolated automatically, and people only patrol and confirm data;
- Reduced labour: necessary intervention points are accepted and kept, but their frequency and duration are pressed down until one operator can attend several machines.
For most assembly shops the second is the realistic target. A closed loop costs not only equipment but coordinated changes in feeding, inspection and logistics, and those returns are hard to see before output and process have stabilised. Reducing labour first, mapping the intervention points, then deciding which of them deserve automation, is the safer order.
Three indicators that tell you whether labour reduction holds
- Interventions per shift: how many times in an eight-hour shift someone has to walk to the machine;
- Average duration per intervention: including walking there and back, not just the seconds of hands-on work;
- Whether the line must stop, and for how long: the most overlooked item. An intervention that takes thirty seconds but requires three minutes of stop-and-reset costs far more than its duration suggests.
Only with all three recorded can you tell whether a line is genuinely "one operator, two machines" or "two machines, half an operator each" - the latter shows up first when staffing is tight.
2. Five intervention points that have to be kept
Replenishment and reloading
Hopper, bowl feeder and reel capacity set the replenishment frequency, and replenishment can hardly be eliminated. With a hopper holding 1,000 screws and eight screws per part, a refill is needed roughly every 125 parts; at a 45-second cycle, 80 parts per hour, that is about every hour and a half, some five times a shift. The frequency is not high; what hits output is whether refilling requires stopping, tools, or re-tuning the bowl.
Changeover and first-off confirmation
Every product change means changing fixtures, calling a program and running a first-off part. This is the longest single intervention and the one that hits the day's output hardest. Changeover should be timed in three parts - fixture swap, program call, first-off fastening and inspection - because a single total tells you nothing about where to improve.
Jam clearing and fault diagnosis
Blocked feed tubes, worn bits, floating screws and cross-threading alarms all need human judgement. Such interventions are unpredictable, so they should not be estimated by average duration; count occurrences per hour and handling time separately, and leave margin for the worst case.
Sampling and reject handling
Sampling by batch or by interval is what makes fastening quality data credible. What happens afterwards matters as much: rejects and good parts must be physically separated, labelled with batch, station and time, so that a mixed lot is not discovered at the customer.
Data review and batch marking
Torque curves, fastening counts and alarm logs are usually reviewed at shift end or handover. This step costs no cycle time, but if it is skipped the traceability chain breaks at the manual link. Turn it into a handover checklist rather than relying on memory.
3. How many machines one operator can attend
"One operator, several machines" is not a slogan; two rough calculations bound it.
The first is time occupancy. Add up each machine's intervention demand per shift and divide by the operator's available hours.
- Eight-hour shift, less about 20 percent for walking, recording and handover: roughly 380 minutes available;
- Per machine per shift: replenishment 5 times at 2.5 minutes, about 13 minutes; faults 8 times at 6 minutes, about 48 minutes; sampling 6 times at 4 minutes, about 24 minutes; changeover twice at 25 minutes, about 50 minutes. Total about 135 minutes;
- 380 divided by 135 is about 2.8 machines.
The second is response time. Once a machine stops, someone must arrive within a defined time, otherwise waiting converts directly into lost output. If arrival within three minutes is required and each machine averages 1.2 faults per hour, attending four machines means nearly five responses per hour - already close to what one person can sustain across a shift.
Take the smaller of the two. In the example above, time occupancy allows about 2.8 machines and response time allows four, so the sensible conclusion is two machines per operator, with a third depending on how concentrated changeovers are. Larger hoppers or fewer changeovers change the result appreciably, so the calculation should be redone against the customer's actual product mix at the proposal stage rather than borrowed as a general rule.
4. Four design measures that lighten frequent interventions
Put the filling port outside the guard
Refilling comes round several times a shift, and once the line stops the loss is counted at full line takt. Placing hopper and reel filling ports outside the guard, or using twin hoppers and twin bowl feeders in alternation, lets replenishment happen without stopping or opening the enclosure - a small investment with a direct return.
Give the low-level alarm lead time
Set the low-level threshold not by "how many screws remain" but by "how long a refill takes". If one refill - walking over, opening the box, resetting - takes about three minutes, the threshold must leave three and a half to four minutes' worth of screws, otherwise the alarm sounds while the machine is already waiting.
Tie alarms to station and cause
A stack light only says "stopped", not "where and why". The HMI should report station number, cause code and suggested action, and keep the last several alarms for shift handover. Diagnosis, not handling, is usually where the time goes.
Keep usable single-step and jog modes
Clearing a fault often means moving one axis or blowing a single screw. Such functions should be callable without removing guards or editing programs, and separated by permission level from automatic running to prevent accidental use.
5. Where the operator stands: walkway and safety
Reducing labour does not lower safety requirements; with more machines per person, the judgement made at each approach needs tighter constraints.
- Separate the operating side from the maintenance side: loading, unloading and replenishment on one side, setup and repair from the other, with guard logic configured separately;
- Interlock doors and light curtains to machine state: no start with a guard open or a curtain broken, and no opening during motion;
- Stop the machine to clear a jam: work in the feed tube or bit area requires stopping, releasing pressure and isolating power first;
- Fix the reset sequence: clear residual screws, retract the bit, home the axes, confirm the part is seated, then allow a start. Written into the work instruction, it is more reliable than individual judgement.
6. Acceptance: four tests that show labour reduction holds
- Continuous run test: four hours at full hopper and normal process, recording actual interventions and their causes against the proposal estimate;
- Jam simulation: induce blocked tube, slipping bit and empty hopper, and time the recovery from stop to restart;
- Changeover timing: from line stop to first-off accepted, timed in sections including fixture change, program call and inspection;
- Real output with one operator on several machines: compare output per person in week one and week five - the gap separates learning from design weakness.
With these four done, how many people a line needs has a basis, and it becomes clear where improvement should be directed.
7. Frequently asked questions
Q1: Must an automated screwdriving line keep an operator?
It depends on whether feeding and inspection can be closed. Where screws must be replenished by hand and faults need human judgement, keeping an operator is necessary. Treating that as a design defect is inaccurate; the point is to leave people only the work that needs people.
Q2: What distinguishes two machines per operator from four?
Not so much walking distance as response time and changeover concentration. Attending four machines, if two or three are changed over on the same day, the operator is occupied in bursts and actual output falls below the average-based estimate.
Q3: How can changeover impact be reduced?
A common approach is to split changeover into parallelisable parts: quick-change, pre-assembled fixtures, programs stored by product number, and first-off inspection off the main line. Based on the company's current project experience, changeover time in high-mix settings can typically be cut to around 60 percent of the original; the exact figure depends on product differences and fixture design.
Q4: Does labour reduction lower the skill requirement?
The opposite. Attending several machines calls for reading alarm causes, judging whether to stop, and resetting in sequence - none of which is simple loading work. Training should be delivered together with the equipment.
Based on the company's current project experience, Cartesian fastening machines hold Z-axis positioning accuracy in the order of ±0.02 mm, with X-Y rapid traverse within 500 mm/s, a common work envelope of 450 × 400 mm, AC 220 V supply, a machine weight of about 200 kg and overall dimensions of roughly 1800 × 800 × 600 mm. Power, air supply and footprint for multi-axis and fully automatic lines must be recalculated against station count and takt. These figures are reference values for standard configurations; specifics depend on the workpiece and the final solution.
— End of Article —
Author: Chisu Automation Technical Team
Published: November 16, 2026
Keywords: automated screwdriving line, labour reduction, manual intervention points, automatic screw fastening machine, automatic assembly line, one operator multiple machines