Testing is cheap and mostly wasted. Most buyers test the wrong things, on the wrong sample, at the wrong time, and then treat a pass as a guarantee. This is a test plan built around the four decisions that actually change.
The minimum set
| Test | Method | Acceptance |
|---|---|---|
| Dimensional change | ISO 6330 + ISO 5077 | <= 5 % after 3 washes |
| Skew | AATCC 179 | <= 5 % |
| Pilling | ISO 12945-2 | >= grade 3-4 at 2000 rubs |
| Colour fastness to washing | ISO 105-C06 | >= grade 4 change, >= 3-4 staining |
| Bursting strength | ISO 13938-1 | report value |
| GSM | ISO 3801 | +/- 5 % of nominal |
Why each test earns its place
Dimensional change. Run to ISO 6330 + ISO 5077, acceptance <= 5 % after 3 washes. This is the test that answers the question the buyer will ask first, and the one where a supplier is most likely to send a development sample rather than production material.
Skew. Run to AATCC 179, acceptance <= 5 %. This is the test that answers the question the buyer will ask first, and the one where a supplier is most likely to send a development sample rather than production material.
Pilling. Run to ISO 12945-2, acceptance >= grade 3-4 at 2000 rubs. This is the test that answers the question the buyer will ask first, and the one where a supplier is most likely to send a development sample rather than production material.
Colour fastness to washing. Run to ISO 105-C06, acceptance >= grade 4 change, >= 3-4 staining. This is the test that answers the question the buyer will ask first, and the one where a supplier is most likely to send a development sample rather than production material.
Sampling for knitted fabric, and where it goes wrong
A single sample from the top of a roll tells you about the top of that roll. For a first order of knitted fabric, take three positions along the run and one sample per colour, then test dimensional change on all of them rather than on a composite.
Condition the samples before testing. Several knitted fabric results move by 10-20 % between a dry laboratory and a humid one, which is why two labs can disagree on the same material without either being wrong.
Reading a knitted fabric report properly
- The measured value for dimensional change, not only the verdict. A pass at <= 5 % after 3 washes is a risk, not a comfort.
- The method number for skew, not the description. A trade name is not a method.
- Conditioning and duration, because both change several results in this family.
- Sampling position and lot number, so the result can be tied to the cartons you received.
- The laboratory name and its accreditation, if the report may be used in a dispute.
The three tests buyers skip and regret
First, the test that reproduces the end use rather than the laboratory condition. For knitted fabric that usually means skew measured after the wash or exposure the product will actually see, because skew / torque does not appear in a conditioned laboratory.
Second, the test on the assembled product. Flat goods can pass every check and still fail at a seam, a fold or a fixing point. Shrinkage beyond claim is usually a construction failure rather than a material one.
Third, retention. Keep one sealed sample per approved lot, labelled with the lot number and the date, for at least two years. It is the only way to know whether weight changed or the conditions did.
Turning test results into a purchase order clause
Write the acceptance value into the PO, name the method, and state what happens on failure: replacement, credit, or rejection at whose cost. A test without a consequence is advisory, and advisory tests get skipped when the schedule is tight.
Full parameter ranges are on the knitted fabric page; the checks that hold those ranges in production are described on the quality control page.
Questions buyers ask before ordering
Can you match a Pantone colour on cotton jersey?
Within dE <= 1.0 on most shades. Very bright greens and some blues are outside the reactive gamut; we will tell you before sampling rather than after.
Do you offer compacted fabric?
Yes, and we recommend it for any garment that will be washed by the end consumer. Compaction adds cost and removes most shrinkage claims.
What is your minimum for a custom colour?
500-800 kg depending on shade depth. Stock white and black can go lower.
How often to re-test
| Stage | What to test | Frequency |
|---|---|---|
| Development sample | All parameters in the specification | Once per development |
| First production lot | All parameters, plus the end-use test | Every lot for the first three |
| Repeat lots | Critical parameters and the one that failed before | Every lot |
| Annual | Full panel, including ageing where relevant | Once a year, or after any process change |
The rule behind the table is simple: test most when you know least. A supplier with three clean lots has earned a lighter regime; a supplier with a process change has not, even after five years.
Retained samples: the cheapest insurance in the order
Keep one sealed sample per approved lot, labelled with the lot number and the date, for at least two years. It costs a shelf. It is also the only way to answer the question that decides most claims: did the material change, or did the conditions?
Without a retained sample, a claim becomes an argument about memory. With one, it becomes a measurement, and measurements get resolved.
For knitted fabric, the sample should include the finish, not just the base material, because skew / torque usually starts in the finish or the additive rather than the substrate.
The test plan for knitted fabric, by order size
| Order | Tests to run | What it establishes |
|---|---|---|
| Sample or trial | Dimensional change | Whether the material can meet the requirement at all |
| First production order | Dimensional change, Skew, Pilling | The baseline every later lot is compared against |
| Repeat orders | Dimensional change plus the parameter that failed last time | Whether the process has drifted |
| Annual or after a process change | Full panel | That nothing has moved in the background |
Testing more on a repeat order is waste; testing less on a first order is a gamble. The table is deliberately asymmetric, because the risk is. For knitted fabric, the one test never worth dropping is dimensional change, because it is the one that predicts skew / torque.
From quotation to specification in seven lines
Leave the tolerance and the method out and every dispute is settled by whoever needs the goods more.
- Structure: jersey / interlock / rib / fleece is the market band. Fix the figure you will test at goods-in, and the instrument you will test it with.
- Composition (100 % cotton / CVC / TC / 100 % polyester): fix the value per shipment, and state who pays for the re-test if a lot lands outside it.
- Weight: 120-320 gsm is not an answer on its own. Ask for the last three lots’ actual readings and the spread behind them.
- Stitch length: agree the test method with the number. Two labs, two methods, two answers, one argument.
- Sample retention: one production sample per lot stays with the mill for twelve months. Claims are settled against it, not against memory.
- Test reports: name the standard as well as the test. A pass on a development sample is not a pass on a production lot.
- Deviation rule: anything outside the agreed bands is reported before shipment, not after arrival.
Questions worth asking before the price is agreed
The questions below are the ones that cannot be answered from a catalogue, which is exactly why they are worth asking.
| Question | Answer that should worry you | Answer that holds up |
|---|---|---|
| Who signs the test report, and against which standard? | “Our QC department signs it.” | A named standard, a named method and a signature that belongs to somebody who can be asked again next year. |
| How do you hold composition inside 100 % cotton / CVC / TC / 100 % polyester? | “That is standard for us, no problem.” | The last three lots’ readings, the instrument used and who signs the report. Shrinkage beyond claim starts here. |
| What happens if stitch length misses the 2.6-3.4 mm requirement? | “It will not happen.” | A re-test clause, the tolerance on the re-test, and the point at which the lot is remade rather than discounted. |
Need a test plan for a first order?
Send the end use, the quantity and the destination market, and we will come back with a specification that meets it, a price at your volume and the test reports that back it up. Samples are free for qualified programmes; we charge the courier only.
Related reading
Sample cut with your values on it · Neutral packing and private label · Consolidation, documents and lead times
