Export Cone Weight Giveaway Control with AI Vision Software
By James C on September 19, 2026
Every cone leaves the winding machine slightly heavier than it needs to be, and for good reason: an underweight cone on an export shipment is a claim, a short-shipment dispute and a conversation with a buyer nobody wants to have. So targets are set with a margin, the margin becomes habit, and the habit is never measured in money. Across a twenty-tonne container a one and a half percent overweight is three hundred kilos of yarn given away free, invoiced to nobody, repeated on every shipment of the year. iFactory's cone-weight vision captures the weight of every cone at packing, shows where the giveaway is being generated, and holds overweight cartons for rework before they are strapped.
AI Vision Cost Control for Yarn Export
Export Cone Weight Giveaway Control with AI Vision Software
Vision AI reads cone and carton weight at the packing station, tracks the distribution by winding machine and drum, and flags overweight cartons for rework before they join the export shipment.
Giveaway Is Invisible Because It Is Never Aggregated
Nobody at the packing table is being careless. A cone thirty grams over nominal looks exactly like a cone on nominal, and thirty grams is well inside what anyone would consider normal winding variation. The loss only becomes visible when you multiply it: by the cones in a carton, the cartons in a container, and the containers in a year. And because weight is usually checked by sampling rather than measured on every cone, the distribution is never seen at all — only a mean that looks acceptable. What sampling hides is the shape: which machine sits consistently high, which drum has drifted since its last service, and how much of the total giveaway comes from a small number of positions that could be corrected this week.
Where the Weight Goes
Plot every cone rather than a sample and the picture is immediate. The distribution sits well to the right of the declared weight, because the spread forces the target upward: the wider the spread, the more margin is needed to keep the left tail clear of a claim.
Cone weight distribution against declared weight
Tightening the spread is what allows the target to come down. A narrower distribution keeps the same protection against an underweight claim while sitting far closer to the weight you actually invoice.
The Loss Has an Address
Once every cone is weighed and tagged with the machine and drum that produced it, giveaway stops being a plant average and becomes a maintenance list. It is common for a small number of positions to generate a disproportionate share of the total.
Giveaway by winding machine
Two machines running the same count on the same order do not produce the same weight. Attribution turns a finance problem into a work order, and the correction usually costs an hour of maintenance.
The Carton Is the Last Place to Catch It
Individual cones vary, but the carton is what gets invoiced and shipped, so the check that matters is the one made as the carton is built. An overweight carton is held for rework; an underweight one never leaves at all.
Carton weight check — export order, 24 cones
Orderexport, declared net per carton
Sourcewinder W4, drums 3 and 7
Carton net against declared+1.9%watch
Cone-to-cone spread in cartonwidewatch
Underweight conesnoneOK
Packing list agreementmatchesOK
Drum 7 trenddrifting high for 3 dayswatch
Carton held for rework: swap the two heaviest cones and it ships on declared weight. Drum 7 is raised as a maintenance job, because holding cartons every shift is treating the symptom.
What the Weight Control Loop Covers
Four things have to work together for giveaway to actually fall, rather than simply being reported.
Weight capture at packing
Vision reads the scale at the packing station, so every cone and carton is recorded without adding a step for the operator.
Machine and drum attribution
Each weight is tagged to its source, turning an average into a specific position that can be adjusted or serviced.
Carton hold and rework
Cartons outside the window are stopped before strapping, when swapping a cone is a ten-second job rather than an unpacking exercise.
Margin reporting
Giveaway is reported as kilograms and as money per shipment and per machine, in the terms finance and the commercial team already use.
What Weight Control Delivers
The recovery is ordinary margin on yarn you are already producing, and it repeats on every container.
Lower
Giveaway per container
kilos recovered on every shipment
Tighter
Weight spread
lets the target come down safely
No
Underweight claims
short cones never leave packing
Named
Machines and drums
drift corrected at the source
Frequently Asked Questions
How much weight giveaway is normal on export yarn?
Plants that have never measured it across every cone are commonly somewhere between one and three percent above declared weight, and are often surprised by their own number. The size depends on winding variation, how conservatively targets were set and how long it has been since the machines were compared against each other. The honest answer is that the plant average is the wrong question: measure the distribution for a week and the number will be specific to your machines, your count range and your packing practice.
Why use vision instead of connecting to the scale directly?
If your scales already expose a data interface, use it, and we will read it. In practice many packing stations run older scales with a display and no output, spread across several lines, and replacing them is a capital project nobody wants. Reading the display with a camera gets the same number into the system in an afternoon, with no change to the scale, the wiring or the way the station is operated.
Will holding cartons slow down packing?
Only at the start, and that is the point. In the first weeks the hold rate reflects how much variation is already there, and most holds are resolved by swapping a cone. As drifting machines and drums get corrected the hold rate falls sharply, because the cartons being built are already inside the window. A plant still holding many cartons after a month has a maintenance problem that was previously being shipped instead of fixed.
Can we reduce the target weight once the spread is tighter?
That is where the saving is realised, and it should be a deliberate decision rather than an automatic one. Once the distribution has been measured and narrowed, you can calculate how much margin is genuinely required to keep the underweight tail clear of your contractual limit, and move the target accordingly. The system reports the remaining risk at any proposed target, so the reduction is made against evidence rather than nerve.
Does it connect to our ERP and packing list?
Yes. Recorded weights are posted against the order, the carton and the batch, so the packing list, the invoice and the shipping documents are built from measured weight rather than from nominal values. That also removes a familiar source of friction at the destination, where a declared weight and an actual weight that disagree turn into a customs question or a buyer query.
Stop Shipping Yarn You Did Not Invoice.
Measure the Giveaway on Your Next Export Order
Give us a week of cone weights, or let us capture them at one packing station. We'll show the distribution, attribute the loss by machine and drum, and put a number on what a tighter target would recover per container.
