A wash tank on Line 1 sat above its learned band for six hours, and the line never stopped
Wash tank 3 on Line 1 left its learned band at and stayed out for six hours. It peaked at in two consecutive hours and still read at nine. Defect rate on the line reached , and Line 1 was the only line outside its own band all day. Inside the Line 1 and SKU-4417 slice the same measure reads . Wash tank 1 on the same line and product moved late, to . Conveyor speed held at and the line kept running through all six hours. The other parameters on the line each held inside its own band all day.
So what? A parameter can sit inside everything the equipment allows and still be wrong for the product being made. That is why every threshold on the line stayed correctly silent. Six hours of drift on one circuit is a controller, a sensor or a valve, and all three can be checked.
537.50K vs 168.91K
the only line outside its band
56 vs 39 °C
two consecutive hours
49 vs 39 °C
six hours out and still out
537.50K vs 496.12K
the slice inside the line
87.08 vs 79.85
above its own, so not the product
47.00K vs 48.00K
inside its running range
Defect rate · every line against its own learned band
One band per line, because each runs different products at different speeds.
The line on each bar is that row’s own expected level. The scale runs from 0 to 600.
Defect rate on Line 1 reached and the line was the only line outside its own band all day. Every band here is learned for that line rather than set once for the plant. That is the only reason one line stands out on a measure all five carry.
Wash tank 3 and wash tank 1, hour by hour · Line 1
The band is what this parameter holds on this line running this product.
Wash tank 3 left its band at , peaked at and was still reading six hours later. Wash tank 1 held its own band through all of it. A long excursion rather than a spike is what points at the controller, the sensor or the valve.
Parameter against hour · the whole circuit set on Line 1
Nineteen parameters are read; these are the five that carry any movement.
| 03:00 | 06:00 | 09:00 | 19:00 | |
|---|---|---|---|---|
| Wash tank 3 | 50 °C | 56 °C | 49 °C | 38 °C |
| Wash tank 1 | 38 °C | 37 °C | 38 °C | 48 °C |
| Curing temperature | in band | in band | in band | in band |
| Conveyor speed | in range | in range | in range | in range |
| Line pressure | in band | in band | in band | in band |
One parameter of nineteen carries the morning, and wash tank 1 on the same line and product moved to in the evening. The other parameters each held inside its own band all day. Maintenance checks one circuit family rather than nineteen parameters, which is the difference the grid makes.
Each against its own level, on the same day and the same line.
| Reading | 31 July, against its own level | What it rules out |
|---|---|---|
| PLC parameter band index · wash tank 3 | 56 °C against its own 39 | nothing — this is the finding |
| Quality yield % · Line 1 and SKU-4417 | 87.08 against its own 79.85 | the product and the recipe |
| Conveyor speed · Line 1 | 47.00K against its own 48.00K | downtime, and a stopped-line artefact |
| PLC parameter band index · the other eighteen | each inside its own band | the line as a whole |
| The equipment's own limit | 56 °C was inside it | a hardware breach — which is why nothing alarmed |
| What is left | the wash tank 3 circuit | a controller, a sensor or a steam valve |
Quality yield on the slice came in at , so the product context is not failing. Conveyor speed held at and the line kept running through all six hours, so this is not downtime. The peak at was inside what the equipment allows, which is why a threshold alarm on the hardware limit never fired.
537.50K vs 168.91K
a full day of production
49 vs 39 °C
and not yet corrected
537.50K vs 496.12K
the product that was running
Defect rate picked the line, not quality
Defect rate on Line 1 reached and the line was the only line outside its own band all day, while quality yield on the slice came in at .
Why it matters: It names which measure did the work, so the next read starts in the same place rather than at the product.
A long excursion, not a spike
Wash tank 3 left its band at , peaked at and was still above its band at six hours later.
Why it matters: A spike is a reading. Six hours is a controller, a sensor or a valve, and all three are checkable while the shift is still on.
The line kept running, which is the point
Conveyor speed held at and the line kept running through all six hours while the parameter drifted.
Why it matters: A defect spike with the line stopped is a downtime artefact. This one happened on running production, so the defects are real output.
What was the controller setpoint on wash tank 3 through those six hours?
Separates a setpoint change from a failing sensor or valve.
How many products ran on Line 1 inside the window?
Sizes the product exposure rather than the parameter one.
Has wash tank 1 moved late on other days?
Tests whether the evening move is the same circuit family or a one-off.
What does wash tank 3 do on Line 2 on the same product?
The control: if it holds there, the circuit is the cause rather than the recipe.
Which parameters sit closest to their learned band right now?
The same read, run before the next excursion.