How to calculate line balancing
Balancing a line answers three questions: what pace demand calls for, how much work goes into each unit, and how to split that work across stations so none of them falls behind the pace.
Takt = available time ÷ demand
Minimum operators = total work content ÷ takt (rounded up)
Balance efficiency = total work content ÷ (stations × bottleneck cycle)
Balance delay = 1 − efficiency
| Term | What it is |
|---|---|
| Takt | the pace demand calls for. See the takt time calculator |
| Total work content | every operation at every station, for one unit |
| Bottleneck cycle | the time of the most loaded station. It sets the real pace of the line |
| Idle time per cycle | stations × bottleneck cycle − total work: the seconds the whole line spends waiting on each unit |
Task times come from a time study. Use the standard time, with rating and allowances, not the raw stopwatch average.
Worked example
The values loaded in the calculator are from a fan assembly line:
- 450 min available per shift, 2 shifts: 54,000 seconds per day
- Demand of 900 units per day
- 17 operations across 5 stations, 204 seconds of work per unit
1. Takt. 54,000 ÷ 900 = 60 seconds per unit.
2. Station loads. S1 = 45 s, S2 = 38 s, S3 = 66 s, S4 = 25 s, S5 = 30 s. S3 is over takt: it is the bottleneck, and the line can only release one unit every 66 seconds.
3. Capacity. 54,000 ÷ 66 = 818 units per day, 82 short of demand.
4. Minimum operators. 204 ÷ 60 = 3.4, rounded up: 4 operators. The line has 5 and still falls short. The problem is not headcount, it is how the work is split.
5. Efficiency. 204 ÷ (5 × 66) = 204 ÷ 330 = 61.8%. Balance delay is 38.2%: on every unit the line spends 126 seconds waiting (330 − 204).
6. Rebalancing. With a target of 95% of takt (57 s per station), the operations are grouped in process order, filling each station as far as it goes:
| Station | Operations | Load |
|---|---|---|
| Station 1 | pick up the base, mount the motor, screw the motor, fetch the fan blade | 54 s |
| Station 2 | fit the blade, tighten the nut, check the clearance, fit the rear guard | 44 s |
| Station 3 | route the cable, wait for the screwdriver, clip the cable, fit the front guard | 51 s |
| Station 4 | test, label, walk to the pallet, pack, close the box | 55 s |
Result: 4 stations instead of 5, a 55 second cycle, efficiency of 204 ÷ (4 × 55) = 92.7% and capacity of 54,000 ÷ 55 = 981 units per day. The line now meets demand with one operator fewer.
How to read the Yamazumi chart
Each column is a station, and each block is a task colored by work type:
- Value adding (VA): changes the product in a way the customer would pay for. Assembling, fastening, packing.
- Necessary non value adding (NNVA): does not change the product but cannot be removed today. Tests, checks, picking up the part.
- Waste (NVA): can be eliminated. Fetching material from far away, waiting for a tool, walking.
The red line is takt. A column above the line is a station holding up the whole line. Very uneven columns mean there is room to move tasks. And the orange blocks (waste) are the first to go: in the example, 23 seconds per unit (fetching the blade, waiting for the screwdriver, walking to the pallet) could become slack without hiring anyone.
How the suggestion works
The suggestion walks through the operations in the order you pasted them and fills each station up to the target. When the next operation does not fit, it opens a new station. It is the same logic as the GBO in TaktZen.
The default target is 95% of takt. Planning stations right at takt (100%) leaves no margin for small stops and operator variation, and many plants plan between 85% and 95%. An operation longer than the target gets a station of its own; the fix is then to split it or improve the method.
Treat the suggestion as a starting point. It does not know that two tasks share a tool, that a station sits on the other side of the line, or that a task is too heavy to repeat all shift. Check it with the people who work the line.
Common mistakes
- Using raw stopwatch averages. Without rating and allowances, every station looks like it has more slack than it does.
- Balancing right to takt. A station at 99% of takt falls behind the first time a screw drops.
- Moving tasks out of sequence. You cannot pack before testing. The suggestion on this page keeps the pasted order.
- Using takt instead of the bottleneck in the efficiency. Some textbooks compute work ÷ (stations × takt). That measures the line as it should be; with the bottleneck cycle you measure the line as it is.
- Stopping at the balance. Check that the cycle covers demand with the production capacity calculator and track real losses with OEE.