Adjustment loss is the time and product lost after the new parts are fitted, while the crew tunes the line until it makes good units at full rate. It is the second half of Nakajima's "setup and adjustment" loss. Measure it from the first restart to stable good running, then reduce it by replacing trial and error with recorded settings and fixed positions.
Setup and adjustment are two different jobs
Setup is the work you can plan: remove the old parts, fit the new ones, load the new material. It takes about the same time on a good day and a bad day.
Adjustment is everything after that: run a few, measure, nudge, run again. Its length can change a lot from one changeover to the next, and that spread is worth explaining.
Nakajima's six big losses put both under one name. Splitting them matters because the fixes have nothing in common.
A worked example
Example: a line with an ideal rate of 40 units a minute has a 70-minute changeover, timed from last good unit to first good unit at full rate. That is the definition from how to measure changeover time.
The first 28 minutes are the swap. Then the crew spends 18 minutes stopped, adjusting between short tries, and 24 minutes running slowly while the line settles.
In those 24 minutes the line makes 480 units, and 360 of them are scrapped. Here is what that costs.
Check the sum. The 120 good units are worth 120 / 40 = 3 minutes, and 28 + 39 + 3 = 70.
Adjustment is 39 of the 67 lost minutes, or 58%, on this hypothetical line. A project aimed only at the swap would miss more than half.
Why your reports understate it
A common habit is to log a changeover as ended when the machine restarts for good. Under that habit, this example shows a 46-minute changeover (28 + 18).
The other 21 minutes do not vanish. They land in performance and quality, far away from the word "changeover", so nobody connects them.
| Part | Minutes | OEE factor | Usually logged as |
|---|---|---|---|
| Swap parts | 28 | Availability | Changeover |
| Stopped tuning | 18 | Availability | Changeover |
| Slow running | 12 | Performance | Nothing |
| Scrap | 9 | Quality | Scrap, no link to the changeover |
The wider idea, that any restart loses output across all three factors, is covered in startup losses. This page is about the tuning that follows a product change.
Where adjustment comes from
Adjustment exists because the first setting was not right. So ask why the crew could not get it right the first time.
| Source | What you see | Fix |
|---|---|---|
| Settings not written down | Crews start from memory or from a guess | A settings sheet per product, kept at the machine |
| No reference on the machine | Handwheels and slots with no scale | Scales, counters, marks, fixed stops |
| Parts that do not repeat | Same settings, different result | Check wear on change parts, locate them with pins |
| Material differs from last time | Fine on one lot, off on the next | Record lot effects, tighten incoming checks |
| Machine not at temperature | Quality drifts as it warms up | Preheat before the stop ends, as external work |
Waiting is the other big one. If the crew is done but quality has not signed off the first piece, the clock runs and the line sits.
How to measure it on the floor
You need four numbers from each changeover, on top of the normal timing sheet.
- Restart time: the first moment the machine runs after the swap.
- Stable time: the moment it holds full rate and good quality, for example for 10 minutes.
- Touches: how many times someone changed a setting between the two.
- Units and scrap: total units made and scrapped in that window.
Touches is the most useful of these. It is easy to count, nobody disputes it, and it should fall as your settings get better.
Compare crews on the same product pair. If one crew needs 3 touches and another needs 14, the first crew knows something that should be on paper.
How to reduce it
The approach is a sequence, and each step makes the next one possible.
Write down every setting while the product is running well.
Add scales or marks, so a setting can be read and repeated.
Change only one thing at a time when tuning, and note it.
Use fixed stops, pins or stored recipes on the settings that never change.
If a new value works better, the sheet changes the same day.
Writing down the settings of a good run is known as centerlining. It works because it gives everyone the same starting point.
The SMED method covers the wider changeover, including its later steps on removing adjustment. Poka-yoke gives ideas for parts that can only go in the right way.
The weekly payoff
Example: suppose the fixes bring that line's adjustment down from 39 minutes to 15 (6 stopped, 6 slow, 3 scrap). That target is an illustration, not a promise.
At 4 changeovers a week, that frees 24 x 4 = 96 minutes. On 2,400 scheduled minutes, that is 4.0% of the week, using nothing but better settings.
The changeover cost calculator turns your own minutes into money. Fewer minutes per change can also change how many changeovers per shift you can afford.
Where tuning projects go wrong
- Ending the clock at restart. The tuning that follows then looks like a speed or scrap problem.
- Tuning several settings at once. When the result improves, nobody knows which change did it, so nothing is learned.
- Calling skill the answer. A gifted setter is good news, but one person's feel does not cover every shift.
- Mixing in running corrections. Nudges for drift later in the run are a separate problem. Most plants log them as short stops or speed loss.
Adjustment shows up as stops, slow running and scrap in the same window. Fabrico takes stops, speed and counts live from the machine, and operators tag the cause at the line, and OEE shows per line and shift. See how Fabrico tracks stops and rate.
What people ask next
Is adjustment loss an availability loss?
Partly. Time spent stopped for tuning hits availability, slow running hits performance and scrap hits quality. The split is shown on the OEE loss tree.
What is the difference between adjustment loss and startup loss?
Adjustment is the tuning needed because the product changed. Startup loss covers any restart, including one after a long stop or a weekend, as in startup losses. Plants differ on where to draw the line, so write down your rule.
How long should adjustment take after a changeover?
We found no neutral benchmark, because it depends on the machine and the product pair. Use your own best run as the bar, and study the gap to your worst. How to measure changeover time shows how to capture the spread.
What is centerlining?
It means recording the settings of a good run and starting from them every time, instead of from memory. It is one of the simplest ways to shorten tuning. The steps above show where it fits.
Does scrap during adjustment count against OEE quality?
Yes, unless your plant has written a different rule. Units made during tuning and then scrapped are defects, and hiding them makes the line look better than it is. See how to calculate OEE quality.
Fabrico is a light MES for OEE. It takes stops, speed and counts from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and operators tag the cause at the line.