Guide · Knitting

Grey fabric rejects in knitting: record every defect by machine and shift

Most knitting floors know their reject kilos. Far fewer know which defect, machine and shift those kilos came from, and that is the information that actually brings rejects down.

By the Accel Safety team · 15 October 2026 · 7 min read

Close-up of white knitted fabric
Photo: engin akyurt on Unsplash

Why reject records matter

In a commission knitting factory, the yarn usually belongs to the buyer and the factory is paid for the fabric it delivers. Every rejected kilo has therefore used the buyer's yarn, machine time and labour without earning anything. If rejects are not recorded properly, three problems follow: the same defects keep coming back, the buyer's yarn balance stops adding up, and sooner or later rejected fabric ends up on an invoice.

A good reject record answers four questions for every rejected roll or piece: what the defect was, where it came from (machine, program and yarn lot), when (date and shift), and how serious it was. Totals alone answer none of these.

Common grey fabric defects, and where they come from

Textbooks list more than twenty knitted fabric faults. On the knitting floor, a short fixed list works better, because a list people can remember is a list people use. The definitions below follow ASTM D3990 terminology as quoted in a 2020 case study from a Bangladeshi circular knitting mill (Hossain and Islam), and the causes come from that study and from S. Natarajan's chapter on knitted fabric defects (INFLIBNET e-PG Pathshala).

DefectWhat it isTypical causes
GoutForeign matter, usually lint or waste, trapped in the fabricDead fibres and fly drifting from neighbouring machines; unclean yarn
Press-offYarn fails to knit and the fabric falls off the needlesYarn-end breakage at a feeder; yarn not entering the needle hooks because of feeder position
HoleYarns damaged enough to leave an openingHigh yarn tension, over- or under-feeding, high take-down force, obstructions in the yarn path, weak places or knots in the yarn
Drop stitchA loop that is not held, leaving a ladderPoor feeder setting, insufficient take-down, wrong stitch cam setting
Miss knitA deviation from the intended knitting patternFaulty positive feed system or wrong feeder setting
Stain (oil, grease)A discoloured area that penetrates the fabricFibre and fluff in the needle tricks combined with excess oil; dirty machines or floors
TuckingOne or more unwanted tuck loopsWeak take-down, dial set too high, wrong course density
Barré and stripesHorizontal or vertical bands in the fabricMixed yarn lots or counts, uneven feeder tension; bent or damaged needles for vertical lines

Two things in this table matter for reporting. First, many causes are machine settings, so the machine number must be on the record. Second, barré and stripes often trace back to mixed yarn lots, so the yarn lot must be on the record too. Our guide to yarn lot traceability explains how to keep that link.

Scoring rejects with the 4-point system

Counting defects is not enough: a pin-sized gout and a long press-off are not the same problem. Most buyers grade fabric with the 4-point system, described in ASTM D5430, which gives each defect penalty points by size. A commonly used allocation is:

Defect size (length or width)Penalty points
Up to 3 inches1
Over 3 and up to 6 inches2
Over 6 and up to 9 inches3
Over 9 inches4
Holes and openings: 1 inch or less / over 1 inch2 / 4

A single defect never gets more than 4 points. The roll's score is then normalised to an area:

Points per 100 square yards = total points × 36 × 100 ÷ (length in yards × width in inches)

Knit fabric is sold by weight, so the roll length is usually not known. It can be calculated from the roll weight, GSM and open width: length (m) = weight (kg) × 1,000 ÷ (GSM × width (in) × 0.0254), then multiplied by 1.0936 for yards. Acceptance limits are set by each buyer; some factories use 40 points per 100 square yards as a pass mark, but always check the buyer's own standard.

A worked example

Take a 50 kg roll of 250 GSM fabric with an open width of 46 inches, inspected with 34 penalty points in total (an example used by OnlineClothingStudy).

  • Length = 50 × 1,000 ÷ (250 × 46 × 0.0254) ≈ 171.2 m, or about 187.2 yards.
  • Points per 100 sq. yd = 34 × 36 × 100 ÷ (187.2 × 46) ≈ 14.2.

Against a 40-point limit this roll passes. But the score alone does not help the floor. The useful record is that, for example, most of those 34 points were holes from one machine on the night shift. That is a tension or take-down problem on a specific machine, and it can be fixed the next morning.

Frequent is not the same as costly

The Bangladesh case study makes this point with real data. The authors analysed three months of inspection records for single jersey, fleece and 1×1 rib fabric, about 55,525 square metres in total, graded with the 4-point method. Gout was the most frequent defect for every fabric type. Ranked by penalty points, the order across all fabrics was gout, press-off, hole, stain, miss knit and tucking. But for 1×1 rib, the most frequent defect was not the most influential one: holes and press-offs carried more points than gout, and the factory's own report had ranked gout first, so its corrective actions were aimed at the wrong defect.

The lesson for any knitting floor is simple: rank defects by penalty points (or rejected kilos), not by how often they are seen, and keep the data long enough to see patterns. The authors also recommend working with the dyeing section, because some grey defects disappear after dyeing while others, such as barré, only show after colouring.

What to record for every reject

FieldWhy
Date and shiftNight-shift and change-over problems only show up by shift.
Machine number and operatorMost causes are settings, needles or feeders on one machine.
Program (order), buyer and yarn lotLinks the reject to the buyer's yarn balance and to lot-related faults.
Roll number and rejected kgKilos are what the buyer's yarn and your bill are measured in.
Defect type from a fixed list"Other" cannot be counted or fixed.
Penalty points or sizeSeparates a cosmetic fault from a roll that fails inspection.
Action takenNeedle change, feeder reset, cleaning: proves the loop was closed.

Record rejects on the same day and on the same line as production. Our guide to the knitting production report shows how that line is built for each machine and shift.

Rejects, yarn and billing

Rejected kilos sit between two accounts. On the yarn side, they consumed the buyer's yarn, so the buyer's balance must show them; how much loss is allowed, and how rejected fabric is treated, should be agreed with each buyer in advance (see our guide to the yarn wastage allowance). On the billing side, the rule is short: bill good kilos delivered, never kilos rejected. When rejects live in a separate notebook, they are easy to miss at billing time, and one wrong invoice can cost more trust than the fabric was worth.

Checklist: grey fabric reject control

#CheckDone?
1A fixed defect list of 8 to 12 types, agreed with QA and the floor☐
2Every reject recorded with machine, shift, operator, program and yarn lot☐
3Rejected kilos weighed, not estimated☐
4Rolls graded with the buyer's points system and acceptance limit☐
5Weekly ranking of defects by points or kilos, not by count☐
6Top machine and defect pair reviewed with maintenance every morning☐
7Rejects shown in the buyer's yarn balance☐
8Rejected kilos blocked from delivery and billing☐

From register to system

A paper register with these columns is a big step up from a single reject total. The hard part comes as machines, shifts and buyers multiply and every reject must stay linked to its program and yarn. AccelKnit, our knitting ERP, was shaped on a working knitting factory: every record sits under one Program ID, rejects are recorded per machine and shift by defect type and are never billed, yarn returns are tracked against a per-buyer wastage allowance, and each machine has its own maintenance log. It runs in English and Bangla. Book a free demo to see it with your own programs, or use our checklist for choosing a knitting ERP to compare systems.

Sources