The plant builds cordless drills and impact drivers in four models, sharing one line. The existing line was a belt conveyor with free-standing benches, installed when the plant ran two models and 600 units a shift. By 2025 it was running four models and 640 units, with overtime covering the rest of the demand.
The constraint, measured rather than assumed
The plant assumed the conveyor was too slow. A time study told a different story: the conveyor ran at 8 m/min, which was fast enough, but the line was 42 metres long because the benches were spaced for storage rather than for work. Operators walked an average of 1.8 km per shift, and the walk was not the only cost: every model change required moving two benches and re-routing the air line.
The second constraint was the functional test. It took 47 seconds against a 32-second takt, and the line stopped while the tester ran. The test was not going to get faster; the test specification had been fixed by the customer.
What we changed
The new line is a 26-metre speed chain conveyor with 14 stations at 1.8 m pitch. The length reduction came from moving material to line-side flow racks and eliminating the storage benches, not from reducing the work content.
- Speed chain with free-flow accumulation, so a blocked station does not halt its neighbours.
- Lift-and-locate units at the four automated stations: two screwdriving stations, one labelling station, one laser marking station.
- Torque capture at both screwdriving stations, with the controller interlocked to the carrier release.
- Two functional test stations in parallel, fed by a diverter after station 12.
- Quick-change fixture tops, one per model, stored on a cart at the line head.
The test station decision
Duplicating the test station cost about 18 percent of the project budget and was the single largest contributor to output. Two testers running in parallel against a 47-second test and a 32-second takt give a theoretical capacity of one unit every 23.5 seconds, which leaves margin for retests.
The alternative, buying a faster tester, was quoted at three times the cost and would have required a new test fixture for each model.
Results after four months
| Metric | Before | After |
|---|---|---|
| Output per shift | 640 | 915 |
| Line length | 42 m | 26 m |
| Walking per operator per shift | 1.8 km | 0.7 km |
| Model changeover | 35 min | 9 min |
| Torque defects per 1,000 | 4.1 | 0.6 |
| Floor space released | - | 96 m2 |
What went wrong during installation
Two things. First, the compressed air line was sized for the old layout, and the two new screwdriving stations starved the existing ones at peak demand. The fix was a receiver tank at the line head, installed three weeks after start-up.
Second, the fixture tops were designed to be swapped in two minutes, but the locating pins were a press fit and needed a mallet. They were changed to a slip fit with a spring detent, which brought the changeover to the intended time.
What transfers to other projects
Three lessons are general. Measure the constraint before designing; the loudest complaint is often not the binding one. Price the bottleneck equipment before the conveyor, because it usually decides the layout. And design changeover for the actual operator, with the actual tools, on the actual cart.
