The Line That Ran Fast Everywhere Except One Place
A customer wanted to increase a packaging line’s output from 60 to 90 parts per minute. They sped up the infeed conveyor, the labeling station, and the case packer. Output stayed at 60. The production manager blamed “slow operators” and “bad morale.” We did a cycle time study. We timed each station over 100 cycles. The infeed took 0.8 s, the labeler 0.9 s, the case packer 1.1 s — but the checkweigher took 2.4 s. That one station was the bottleneck. Everything else could run faster; the line could not go faster than the slowest station. Speeding up the others just created piles of work-in-progress. We optimized the checkweigher (faster belt, better filtering) and dropped it to 1.2 s. The line jumped to 85 parts per minute — no operator change, just finding the bottleneck.
Cycle time analysis is the discipline of finding the slowest station and fixing it. Everything else is noise. Here is how to do it.
The Bottleneck Rule
The throughput of a line is set by the slowest station — the bottleneck. No matter how fast the other stations run, they cannot outpace the slowest one. Work-in-progress piles up between the fast stations and the bottleneck. The line’s capacity is the bottleneck’s capacity, not the sum of all stations.
Speeding up a non-bottleneck station does nothing for output. It only creates inventory. The customer who sped up the infeed conveyor made a pile of parts at the checkweigher, not more finished cases.
How to Find the Bottleneck
Time each station over at least 50 cycles. Use a stopwatch (or, better, a PLC trace). Record the cycle time for each station, and the variation (the slowest cycle). The station with the longest average cycle time is the bottleneck. The station with the largest variation is the next bottleneck to fix — even if its average is not the longest, the outliers slow the line.
Do not guess. “That station looks slow” is not a bottleneck analysis. The checkweigher looked fine (the conveyor ran fast), but its settle time after weighing was 1.4 s. The stopwatch exposed it.
The bottleneck rule: The line runs at the speed of the slowest station. Find it by timing every station over 50 cycles. Fix that one. The other stations can wait. Speeding up a non-bottleneck only creates piles.
Throughput vs Cycle Time: The Difference
Cycle time is the time a station takes per part. Throughput is parts per minute. They are related but not the same. A station with a 2 s cycle produces 30 parts per minute. A station with a 1 s cycle but 50% downtime produces 30 parts per minute too. The bottleneck may not be the slowest cycle — it may be the one that stops most often. Track both cycle time and uptime. A station that cycles fast but jams every 10 parts is the bottleneck.
The Buffer Between Stations
A buffer (a queue of parts) between a fast station and a slow one absorbs the variation. Without a buffer, the fast station waits for the slow one (blocked), and the slow one waits for the fast one (starved). A small buffer (3–5 parts) decouples the stations and lets the line run near the bottleneck’s true capacity. The customer’s line had no buffer — the infeed jammed the checkweigher, and the checkweigher waited for the infeed. A 5-part buffer would have hidden the small variations.
A Cycle Time Study Template
| Station | Avg Cycle (s) | Slowest (s) | Downtime % | Capacity (ppm) |
|---|---|---|---|---|
| Infeed | 0.8 | 1.1 | 2 | 73 |
| Labeler | 0.9 | 1.2 | 3 | 65 |
| Checkweigher | 2.4 | 3.1 | 5 | 24 |
| Case packer | 1.1 | 1.5 | 4 | 52 |
The checkweigher at 24 ppm is the bottleneck. The line cannot exceed 24 ppm. Fix it, and the next bottleneck appears (the case packer at 52 ppm). You optimize one at a time.
A Bottleneck Hunt Checklist
- Time every station over 50 cycles (stopwatch or PLC trace).
- Record the average and the slowest cycle per station.
- Record the uptime % per station (how often it stops).
- Identify the station with the lowest effective capacity (ppm).
- That is the bottleneck — fix it first.
- After fixing, re-time the line; find the next bottleneck.
- Add a 3–5 part buffer between stations to absorb variation.
- Do not speed up non-bottleneck stations (it creates piles).
- Track the line’s actual output (parts/hour), not the design output.
- Re-study after any product or recipe change.
The Bottom Line
The line that ran fast everywhere except the checkweigher was a bottleneck hiding in plain sight. Output is set by the slowest station, not by the fastest. Time every station, find the one with the lowest effective capacity, fix it, then move to the next. The 60 ppm line that jumped to 85 ppm was not a morale problem — it was a stopwatch away.
