Shoulders That Collapse After Two Wears: Spec the Cloth

A coat that holds its line on a hanger and folds at the shoulder after two wears is telling you one thing: the cloth was chosen for weight and handle, but never specified for bending recovery. Structure in a tailored coat is not grams per square metre. It is the cloth’s resistance to bending, how much of that resistance survives a few thousand flex cycles, and how well the shoulder composite of shell, interlining and canvas returns after being crushed. All three are properties a mill can state on paper, and all three can be checked on a sixty metre lot without a laboratory. What your order was missing is not a better fabric test. It is three lines in the specification that name these properties and tie them to the sample you approved.

What structure actually is before you write it down

Ask five people what makes a coating cloth “structured” and you will get weight, wool content, weave density and hand feel. Weight matters, but it is the weakest of the set. Two cloths at the same grams per square metre can behave completely differently over a shoulder, because the one property that holds a shoulder line is bending behaviour: how the cloth resists folding, and whether it comes back.

The first component is bending length, the distance a strip of cloth extends before it folds under its own weight. Laboratories measure it with the cantilever method described in ASTM D1388, the stiffness test named in the order, and the mill can state a value for the production cloth, warp and weft, after finishing. A short bending length means the cloth folds readily; a long one means it holds a line. Instrumented systems such as KES and FAST add the second component, which is recovery. Every cloth bends, but a structured cloth stores energy and returns. A soft cloth bends and stays bent, and that difference, called hysteresis, is what separates a shoulder that recovers overnight from one that memorises every wear.

The third component is compression recovery, and it is the one designers never specify. The shoulder of a worn coat is not only bent, it is pressed: bag straps, seat backs, hanger pressure. A lofty tweed can look thick, feel springy in the hand, and still flatten permanently where the pressure sits, because crush resistance and bending resistance are different mechanisms.

Wool adds a complication of its own. It takes moisture in and gives it off with the weather, and its mechanical behaviour moves with it, which is why the same coat can read firmer in a dry week and softer in a damp one. The Internationale Wolltextil-Organisation publishes the conventions the wool trade uses to describe and test fibre behaviour, and they are the reason a property claim made in one mill’s language can be read in another’s.

Why the showroom coat passes and the worn coat fails

A hanger is a static test. It loads the shoulder once, in one direction, at room temperature, and then leaves the coat alone. Two wears load it hundreds of times, in every direction, with moisture and heat from a body added. Pressing in the factory can set a shape into cloth that has poor recovery, and that pressed shape photographs beautifully in the showroom and fades over the first week of real use. The table below is how the failure presents, where it usually sits, and the first check worth making on a returned coat.

The last row matters because it is the most common misdiagnosis. A coat that has shrunk across the back reads as a collapsed coat, and no bending specification would have saved it. Work through the returns in an order, not all at once.

Working the returned coats in sequence

Three customers returned the same style, so you hold three pieces of evidence. Read them in the order the layers were assembled, because each layer clears or convicts the next.

Read the interlining and canvas first

The shoulder is a composite: shell cloth, fused interlining or sewn canvas, wadding, and the sleeve head. Delamination, resin pockets and a fusing press that ran cold all present as a shoulder that gives up, and they are cheaper to find than a cloth fault. Pinch the front panel between finger and thumb; you should feel a distinct layer sliding inside the shell. Then check whether the fused area still bonds at the armhole seam. If the composite has failed, the cloth never gets its chance to testify, and the trial that should have caught this is the one described in testing the actual cloth with the actual interlining before the bulk was fused.

Read the making next

A shoulder pad that has shifted, a sleeve head wadding trimmed too thin, a pressing buck that was hotter than approved, all produce the same customer sentence: the shoulders collapsed. Open one return at the shoulder seam and photograph what you find before anyone alters it. Compare the making against the approved sample garment, if one exists, not against memory.

Then test the cloth, and only then

Once assembly is cleared, the cloth faces its own comparison, and it needs a witness. Cut a strip from the return at a damaged shoulder area if the customer has agreed, or from an undamaged panel, and lay it beside the retained swatch you approved. Fold both over the same table edge and let them drop; bend both and release. If the returned cloth is visibly limper than the swatch, the lot drifted, and the specification lines below are what would have caught it. If they behave the same, the cloth did what you approved, and the fix is in pattern, interlining or making.

Checking a sixty metre lot without a laboratory

A lab report on a sixty metre lot costs real money against the margin, and the mill’s own sheet only covers the cloth it chose to test. What a small brand can run instead is a comparative protocol, because you are not trying to measure the cloth in absolute terms. You are trying to find out whether this lot behaves like the sample you approved, and relative comparisons are something a cutting table can do well.

1. Cut three strips across the width, one from the head, one from the middle and one from the tail of the roll, each long enough to fold over a table edge with length to spare.
2. Lay each strip over the edge beside the retained approval swatch and photograph where the tips settle. The production strips should hang within a finger’s width of the swatch. A tail that folds shorter than the head is a finished-cloth drift that no single middle sample would show.
3. Press a thumb into each strip for a fixed ten count, release, and photograph at one minute. Recovery speed is visible: a crush mark that lingers on the production strip but not the swatch is the failure that ends up at the shoulder.
4. Run the same fold-and-release on a small sandwich of production cloth fused with the production interlining, because the composite is what the coat is actually made of.
5. Weigh a measured piece against the specification, so a weight drop is caught as a weight drop and not confused with a bending change; the method is set out in naming the method before you argue about roll weight.

Judge the results with the swatch in the room and the same person doing every comparison. Photographs help, but the discipline that makes this protocol work is boring: same table, same person, same day, swatch present. Record what was compared and what passed, because that record is what turns a hunch into a claim if the lot fails later.

The three lines to put in the order

None of the above works retroactively. The protocol has teeth only when the order confirmation asks the mill to state the properties, because a stated property can drift and be caught, while an unstated one was never promised. Three lines cover the mechanism of a collapsed shoulder.

Notice that none of these lines demands a universal number. Absolute stiffness values are meaningless without the pattern, the pad and the canvas that surround the cloth, which is why the reference that matters is the sample you already approved. Asking for “no lower than the approved reference, by a named method” is a request a mill engineer can answer in an afternoon, and it converts a vague hope of structure into something that can drift, be measured and be argued.

Keeping the second lot in the same behaviour

The reorder is where this pays, because bending behaviour is the first property to move when a yarn lot or a finish route changes, and the last property anyone re-tests. Weight gets checked on arrival by habit. Recovery gets discovered by your customers in November. When the reorder is approved, put the three lines from the table into the re-approval itself, and ask the mill to run the same comparative reads on the first production metres against the retained swatch before the lot ships. Where the sample basis has changed, for instance a new yarn lot woven to the same recipe, the approval conversation is the same one as approving bulk from a rebuilt sample, covered in what to compare before you approve.

A shoulder that survives two wears is not luck, and it is not only the tailor’s craft. It is three properties, written down, checked against a swatch, and carried across lots. The coat holds the line the specification drew.