Same fabric

The factory said the fabric hadn't changed. Every test agreed. The garment didn't.

We built a new silhouette for SquadGEAR, our line of sublimated uniforms. The tech pack called out a fabric we had already run in other silhouettes. Proven material, known supplier. The first samples came back right.

Then the size set arrived, the last checkpoint before production, and the stretch was wrong. In certain areas of the garment, the fabric wasn't stretching and recovering the way it had in the original samples.

Nothing had changed on paper. That was the problem.

Rule out the input first

The first question in any materials failure is whether you got the material you specified. We asked the factory. The factory said it was the same fabric.

A factory's assurance is a claim, not a verification. So we asked for an unprinted cutting off the roll and tested it ourselves. It stretched and recovered exactly like the original samples.

Then the punch. I cut a weight sample from the finished garment and one from the raw cutting. They matched. That rules out more than it looks. If the knitter had changed the density or the fiber composition, the weight would have moved. If the press had shrunk or compacted the fabric, the finished weight would have come up. Neither happened.

Then the pick glass. Same knit construction.

Three independent checks — raw stretch, weight, construction — all clean. By every measure we had, the factory was telling the truth. It was the same fabric.

Test it the way it gets used

Every one of those tests had one thing in common: they measured the fabric at rest. A sublimated uniform doesn't live at rest. Before it becomes a garment, it goes through a heat press at 410°F for 50 seconds.

So I ran one more test. I took the unprinted cutting, pressed half of it at 410°F for 50 seconds — production temperature and dwell — and left the other half alone. One piece of fabric, one variable, a built-in control.

The unpressed half stretched and recovered correctly. The pressed half didn't.

The fabric was fine until it met the process.

What changed

The knitter had switched the elastomeric yarn — the spandex that gives the fabric its stretch and recovery — most likely to a lower-grade yarn with a lower melt point, and probably a cheaper one. The knitter told no one. Not us, not the factory. When we traced it back, the knitter acted as if nothing had changed.

Elastane is the most heat-sensitive fiber in a performance knit, and a lower melt point is exactly the kind of difference that doesn't show at room temperature. There, the new yarn behaved like the old one: same fiber content, same weight, same construction, same stretch off the roll. Under sublimation heat, it didn't hold.

That's why every static test passed. None of them were wrong. They measured properties that hadn't changed and missed the one that had.

The spec only covers what it names

I've argued before that the tech pack is the contract. It is. But a contract only binds what it names. Ours named the fiber type and the percentage. The new fabric met both. Spandex at the specified percentage is still spandex at the specified percentage, whatever its grade. What the tech pack didn't say was how the fabric had to perform after it came out of the press.

That gap is where the substitution lived. The knitter swapped the yarn, the fabric still matched the spec on paper, and the factory received what it had always ordered. Nothing in the documentation was violated. The change existed only in how the fabric behaved under heat.

That's the general pattern. The supplier you deal with is rarely the one who makes the change. It's their supplier, usually solving a cost or availability problem you'll never hear about. "Same fabric" is true at the level of the product name and false at the level of the yarn.

What we do now

We already tested incoming fabric after heat — for shrinkage. Shrinkage is the post-print failure everyone knows to check. Stretch and recovery weren't on the list, because the fabric had never failed that way.

Now incoming raw material is verified for post-print performance, not just shrinkage. Before a roll goes to print, a cutting goes through the press at production settings and gets tested for the properties the garment depends on. It's a short test. It would have caught this at the roll instead of at the size set.

What made this dangerous is that the defect is latent. A finished uniform made from this fabric looks right, measures right, and passes end-of-line inspection. Nothing on the production floor would have flagged it. It shows up only when the garment is worn, stretched, and expected to recover — on the field, with the customer.

The size set caught it because we compared stretch and recovery against the original samples and noticed the difference. That's not a control. It's attention, and attention doesn't scale. A last checkpoint shouldn't be the only place a problem can surface.

Test the material in the state it will be used. A fabric that passes at rest has only proven it can sit on a shelf.