English is currently available.
Manufacturing & Production Engineering DISCUSSION
I am planning a manual assembly line for a small gearbox. The shift is 8 hours with 30 minutes of breaks, and the customer needs 360 units per shift. The assembly is split into 9 tasks whose times add up to 410 seconds, the longest single task being 62 seconds.
How do I get from these figures to a takt time and a number of workstations? I also see "cycle time" and "takt time" used as if they were the same thing, and I am not sure they are.
Takt time is set by the customer, not by the line: available working time divided by demand. Here the available time is 480 - 30 = 450 minutes, or 27,000 seconds, so takt = 27,000 / 360 = 75 seconds per unit. One finished gearbox has to leave the line every 75 seconds on average.
Cycle time is what a station or the line actually achieves. A line meets demand only if its slowest station has a cycle time at or below takt, so the two numbers are compared with each other but they are not the same thing.
The theoretical minimum number of stations is the total work content divided by takt, rounded up: 410 / 75 = 5.47, so 6 stations. Balance efficiency is then 410 / (6 × 75) = 91 percent, meaning about 9 percent of the paid station time is idle by design.
Six is a lower bound only. Tasks cannot be split freely and they have precedence rules (the shaft goes in before the cover), so the real assignment may need a seventh station. Your longest task of 62 s fits inside 75 s, which is good: any task longer than takt would need two parallel stations or would have to be divided.
For nine tasks you can balance by hand. Draw the precedence diagram, then fill station 1 with tasks whose predecessors are done, choosing the longest one that still fits in the remaining time, and move to the next station when nothing fits. That is the largest-candidate rule. The ranked positional weight method is similar but sorts tasks by their own time plus the time of everything that follows them, which tends to pull the tasks with long chains behind them to the front.
Try both and keep the result with fewer stations and the smoother spread of station times.
Adding a caution to the 75 s figure: it assumes the line runs for every one of those 27,000 seconds with no rejects. Real lines lose time to short stops, material shortages and rework. A common approach is to plan to a target cycle time below takt. If you expect 90 percent uptime, plan for 75 × 0.9 = 67.5 s. The minimum station count then becomes 410 / 67.5 = 6.07, which rounds up to 7, so the allowance costs you a station in this case.
It is worth knowing that before the layout is fixed rather than finding it out through daily overtime.
Once it is running, check it with a stopwatch rather than the spreadsheet. Time each station over 20 or more cycles and plot the results against the takt line. The station closest to takt is the bottleneck and sets the line output; stations far below it are where operators wait. Moving one small task from the bottleneck to a lighter neighbour often gains more than any other change.
Also recalculate when demand changes. At 300 units per shift takt becomes 90 s, 410 / 90 = 4.6, and the same work can be rebalanced across 5 stations.
