Strength, hairiness and Classimat

The situation

The second page of a yarn report is not more of the same. The first page described the thread as a whole. This page is about where it breaks, the fibre ends that stand out of it, and the lumps along its length. Those are the things that stop machines, and the things a customer eventually sees. The second half of the page leads to a decision most people treat as a matter of taste. It is really arithmetic.

It is 11:20, the same morning, with the same two certificates. Page one was statistics about the whole yarn. Page two is about the parts of it that break, stick out, and stop machines. It is the page that decides how the knitting floor will run next week.

Strength is three numbers, and they are not interchangeable

RKM is breaking length, in kilometres: the length of yarn whose own weight would snap it. Tenacity in cN/tex is the same measurement in different clothes. Divide RKM by about 1.02 and you have it, so an RKM of 17.8 is a tenacity of 17.5 cN/tex. Do not mix the two up in an argument with a spinner. He will correct you, and from that moment you are the person who does not read certificates.

CSP is the count strength product. A 120-yard lea of yarn is wound on a measuring reel and broken on a lea tester, and the breaking load in pounds is multiplied by the count in Ne. It is an old test. It destroys more yarn than a single-end test and tells you less. But it is still the number two mill managers quote at each other on the phone, so learn to convert it. CSP 2,540 on a 30.2 Ne yarn is a lea strength of 84.1 pounds.

Elongation at break is the percentage the yarn stretches before it snaps. On a knitting machine this matters more than tenacity, and most people miss it. A latch needle pulls a loop through in a few milliseconds. What the yarn meets is a sudden stretch, not a slow pull. A strong yarn with low elongation is brittle at the needle. It gives you yarn breaks, machine stops, and a repair mark in the fabric at every one of them.

CV of single-end strength is the number that predicts the breaks. A mean RKM inside the specification, with a strength CV of 10 percent, means the lot has a weak tail — and the weak tail is what fails at the needle. It is never the average bit of yarn that breaks.

On strength, the percentile runs the other way. This catches people out every season. Less irregularity is better, so on page one a low percentile means a good yarn, and the 25 percent line is a low number. More strength is better, so here the 25 percent line is a high number: 18.4 RKM is the value only a quarter of world production beats. A-7741 at 17.8 sits between the 25 and 50 percent lines. A-7768 at 16.2 sits between 50 and 75, with a strength CV of 9.8 that is nearly at the 75 percent line. That combination — a middling average, a wide spread, low elongation — is a yarn that will stop machines.

Twist, and the T-shirt that turns on the customer

Twist multiplier is turns per inch divided by the square root of the count. It is the single number that says what the yarn was spun for. Weaving yarn is usually spun at 4.0 to 4.4, because a warp end has to survive a loom. Yarn for single jersey knitting wants 3.4 to 3.8: enough twist to hold the fibres together, not so much that the yarn stores a leftover turning force inside itself.

That stored force does not disappear when the yarn is knitted. It sits in the loop. The first time the garment is washed, it lets go: the vertical stitch lines lean over, the side seam walks round the body, and the buyer's test report calls it spirality. Most buyers allow 5 percent after three home washes. A-7741 has a twist multiplier of 3.9 — above the knitting range. On a single jersey, 3.9 is worth a wash test on a sample tube before you knit four thousand metres of it.

Hairiness: what stands out of the yarn

H is the total length of protruding fibre ends, in centimetres, per centimetre of yarn. H 4.1 means four point one centimetres of fibre sticking out of every centimetre. Hairiness drives three things a buyer will eventually write to you about. First, pilling — the protruding ends are what roll up into little balls on the fabric. Second, hand — a hairy yarn feels warmer and less crisp. Third, fly — in a knitting hall, those loose ends come off and settle on the next machine.

CV of hairiness is the one that makes stripes. H itself is an average. A fabric knitted from cones that all sit at H 4.9 looks perfectly even. A fabric knitted from some cones at 4.1 and some at 4.9 does not, because hairiness changes how light comes off the surface, and how the dyed colour reads. A working rule: an H difference of about 0.5 between cones feeding the same machine is visible in a pale, fine-gauge knit. The gap between these two lots is 0.8.

Put that beside the count gap from lesson one and see what you are holding: two separate causes that produce the same visible band. If the fabric stripes, you will not be able to tell by looking which one did it. Neither will the spinner.

Classimat: the individual faults that statistics hide

Everything so far described the yarn as a population. Classimat does the opposite. It runs long lengths of yarn — 100 km is the usual test — and classifies every objectionable fault one by one. It sorts each fault by two things: how much thicker or thinner it is than the yarn, and how long it is.

The classic grid has 23 classes. Short thick places get a letter for length and a number for mass. A is up to 1 cm, B is 1 to 2 cm, C is 2 to 4 cm, D is 4 to 8 cm. 1 is plus 100 to 150 percent, 2 is plus 150 to 250, 3 is plus 250 to 400, 4 is above plus 400. That makes sixteen classes. Then E, F, G and H are long thick places, 8 cm and longer. I1, I2 and I3 are long thin places, 8 cm and longer, at minus 30 to minus 75 percent.

Setting the clearer is arithmetic, not taste

The electronic yarn clearer sits on the winding machine. It watches every centimetre, cuts out whatever it is told to cut, and the two cut ends are joined again with a pneumatic splice — a joint made by tangling the fibres together with a blast of air. So every cut is a splice, and the setting is a trade between two costs that move in opposite directions.

Add up the "yes" rows. On A-7741 that is 68 cuts per 100 km. On A-7768 it is 215. A 1.9 kg cone of Ne 30 holds about 96 km of yarn, so cuts per 100 km is very nearly cuts per cone. That means a splice roughly every 1,400 metres on the good lot — and every 450 metres on the bad one.

Now look at what you decided not to cut. If you also cleared the A1, A2, B1 and B2 classes out of A-7768, the count would go from 215 to about 2,790 cuts per cone: a splice every 35 metres. That is not a cleaner fabric. It is a different product. Winding time collapses, the number of splices in the finished garment climbs through the roof — and the faults you removed were lumps under 2 cm at plus 100 to 250 percent, in a 150 gsm jersey, which no consumer has ever seen.

The cost lands on both sides. For the spinner: an automatic winder running at 1,200 m/min takes about 80 minutes to wind a 96 km cone. At roughly 6 seconds per cut-and-splice cycle, 215 cuts adds 21 minutes — 27 percent more winding time on that lot. That is exactly why the person you are negotiating with does not want to clear harder.

For the knitter, the cost is stops. On the same machines, in the same week, A-7741 ran at 1.1 stops per 100 kg and A-7768 at 3.2. Across the 1,080 kg that will be knitted in this course, that is 12 stops against 35, at four to six minutes each. And every stop leaves a repair mark that a fabric inspector points at.

Act as a senior yarn technologist auditing a delivery of yarn against its test certificates. I will paste the certificates. Delivery facts: spinner [NAME], purchase order [NUMBER], delivery note [NUMBER], total quantity [KG], number of lots and the weight of each [LIST], price [PER KG], date delivered [DATE]. The yarn is for: [FABRIC CONSTRUCTION], [GSM], machine gauge [GAUGE] with [NUMBER] feeders, shades [LIST]. Here are the certificates, lot by lot: [PASTE EVERY FIGURE, INCLUDING THE TEST CONDITIONS AND SENSITIVITIES]. Do the following. First, restate each lot's figures against the Uster Statistics percentile lines for this yarn type and count, and say for each row which band the lot falls in and whether that is good enough for THIS fabric rather than in general. Second, tell me every figure the certificate does NOT report that it should, and what each omission hides. Third — and this is the one I care most about — compare the LOTS with each other, line by line, and tell me the largest gap between them in count, hairiness, twist and imperfections, converted into a percentage difference, with a plain statement of what each gap will look like in the finished cloth. Fourth, tell me whether these lots can safely be knitted on the same machine, and if not, exactly what instruction has to go on the knitting programme and the store issue slip. Fifth, list the tests I should run myself on arrival, in the order I should run them, with the time each takes. Do not tell me a lot is acceptable because it passes its own specification; tell me where it sits against world production and against my fabric.