How Much Scrap Does a Coil Line Make Before Good Parts?

Count the Strip Between the Coil and the First Good Part

Four lengths sit between a fresh coil and a part you can ship, and only one of them is usually written down.

Threading length. The strip has to travel from the uncoiler through the straightening and feeding section to the first station. On a mid-size line that is 2.0 to 3.5 m, and it becomes scrap.

Progression length. Every station needs one pitch before the strip reaches the last station. A 14-station die at 60 mm pitch consumes 840 mm before a complete part exists.

Pilot and setting length. The first parts are used to set pilot release timing, check for shaving, and adjust the loop. Plan on 10 to 25 strokes of development.

Inspection length. First-off approval takes parts off the line. On an automotive job that can be five pieces; on a simple bracket it is one.

Add them and the startup block is a fixed cost per coil, not per part. That is the key distinction, because it means coil weight changes the percentage without changing the absolute loss.

Run the Arithmetic on a 1.2 mm Bracket

Take a 320 mm wide coil of 1.2 mm SPCC, 1,500 kg, feeding a 14-station progressive die that makes a 0.42 kg bracket at 55 mm pitch.

The coil holds about 497 m of strip, which is roughly 9,000 pitches. Threading plus progression plus setting plus inspection comes to about 5.3 m, or 96 pitches.

So the startup block costs 96 pitches out of 9,000, about 1.07 percent of the coil. Weighted for the two parts per stroke that never form completely, the practical figure lands nearer 1.3 percent.

Now change one variable. Drop to a 700 kg coil and the strip length halves while the 5.3 m block stays fixed, so the loss doubles to roughly 2.6 percent.

That is the whole argument for heavier coils, and it has nothing to do with machine quality. It is geometry: a fixed loss spread over a longer run.

Set it against handling. A 1,500 kg coil needs a coil car or a crane and a floor that carries the load, so the saving is real only where handling already exists.

Where the Yield Leaks After Startup

Startup is the visible loss. Four others run quietly through the shift, and on most lines they add up to more than the thread-through ever did.

The second row is the one buyers miss. On a 14-station die, the last 800 mm of a coil often cannot complete a part, so it is scrapped as a remnant even though the strip is perfect.

That is where coil joining earns its place, and also where it introduces a new failure mode: a weld thicker than the strip will mark a roll or jam a pilot.

Die Layout Moves the Number More Than the Machine Does

Two dies making the same bracket can differ by a full percentage point of yield, and the difference is decided long before the coil line is specified.

Station count. Every added station adds one pitch of progression scrap per coil. Sixteen stations instead of twelve is 220 mm more per coil on a 55 mm pitch.

Pitch. A longer pitch buys tool strength and costs yield, because progression scrap scales linearly with it.

Web and carrier width. The skeleton between parts is scrap by design. Rotating the blank 90 degrees can change the number of parts per metre by 8 to 12 percent on a rectangular part.

Scrap return. A die that cuts scrap into short pieces moves it out of the way and sells it at a better rate than long tangled web.

None of these are machine decisions, which is exactly why a coil line quotation cannot promise a yield figure. Ask the die designer for the strip layout and compute the loss yourself.

Line capacities that bound the calculation are listed on the coil handling product index.

Set the Acceptance Threshold Before You Sign

A yield clause is only useful if it names a number and a test. Both are easy to write, and both are usually left out.

State the expected scrap rate as a band, not a single figure, and define it as weight of good parts divided by weight of coil consumed over a full coil, not a sample.

For a 14-station bracket on 1,500 kg coils, 2.5 to 3.5 percent total is a defensible band. Under 2 percent usually means someone is counting only the thread-through.

Add the conditions that void the band: coil width tolerance, material grade, and whether the die is new or rebuilt. A yield promise measured on a worn die is not a yield promise.

Then agree the measurement. Weigh the scrap container per coil for the first week. It is crude and it is the only method that survives an argument.

What the Arithmetic Cannot See

The model above treats scrap as length, which is honest for a rectangular blank and wrong for several real cases.

Metallurgical loss. When a part is rejected for a crack or an orange-peel surface, the strip was consumed and no length accounting catches it. Those rejects live in the quality data, not the yield data.

Test and approval runs. A new die may consume two coils before anyone agrees it is good. That belongs in tooling cost, not production yield, and mixing the two hides a real number.

Width tolerance. A coil 2 mm over width yields more parts per metre but may not fit the guides. The arithmetic rewards a dimension the machine cannot accept.

Short runs. Below about 300 strokes, the fixed startup block dominates and the percentage becomes meaningless. Quote those jobs per piece, not per yield.

What to Measure on the First Three Coils

Three coils is enough to confirm or kill the model, provided the crew records the same four things each time.

• Strokes from thread-through to the first part that passes inspection.
• Strokes lost to pilot faults or feed alarms during the run.
• Remnant length left on the mandrel at coil end.
• Scrap weight per coil, weighed rather than estimated.

Repeat the same four readings on the second and third coil, and the spread tells you whether you are looking at a design loss or a setup problem.

• A stable spread of 0.2 percentage points means the loss is structural.
• A spread above 1 point points at the loop, the pilot timing, or the operator.

Our own field engineers use this three-coil test during commissioning, because it separates a machine problem from a die problem before anyone starts swapping parts.