The fifth reorder drifts because the reference was never written down, not because the line got careless. Eight variables move between runs: dye lot, fabric batch, machine, operator, trim, print, foam and packing. Lock each in a spec file at run one, keep a sealed gold sample, and require a written change note before any substitution. MOQ 500 pcs, 300 on a repeat.
The fifth reorder drifts because the reference was never written down, not because the line got careless. Eight variables move between runs: dye lot, fabric batch, machine, operator, trim and hardware, print and embroidery, foam and stiffener, and packing. Lock each one in a spec file at run one, keep a sealed gold sample plus a retained unit from every run, and require a written change note before any substitution. MOQ 500 pcs, 300 on a repeat, bulk 35-50 days.
What Actually Drifts Between Run One and Run Five?
Drift is treated as a quality problem and it is almost always a documentation problem. The line that made run one is not careless by run five; it is working from a newer fabric lot, a partially substituted trim set and a verbal instruction that replaced a written one somewhere around run three. Nothing was decided badly. Nothing was written down, so nothing could be checked. The bag moved, and by the time somebody noticed, the reference needed to prove it had moved was gone.
The useful way to think about drift is as eight independent variables, each with its own owner and its own control. They are independent in the sense that fixing one does nothing for the others: a locked dye lot does not stop a substituted buckle, and a retained gold sample does not stop a new screen changing logo placement. Consistency on run five is the product of eight controls, which means it fails at the weakest one and not at the average of them.
| Variable | What actually changes | The symptom on the shelf | The control that works |
|---|---|---|---|
| Dye lot | A new bath, a fresh recipe weigh-out, a different water temperature and time | Two cartons in one shipment do not match; run five sits half a shade from run one | Record the dye-lot number at approval and re-book the same recipe by number |
| Fabric batch | A different weaving batch with a different pick count or coating weight | Hand feel and stiffness differ; the bag drapes differently when it is empty | Lock the mill and the fabric code, and file a weighed swatch from every new lot |
| Machine | A different needle plate, feed dog or machine class on a busy week | Stitch length and seam pucker change; the stitch count per inch is not the same | Name the machine class and the stitch specification in the spec file |
| Operator and line | A different crew, a different line, a different daily target | Topstitch alignment and corner shaping vary within a single run | Keep the style on the same line and keep the first-off check every run |
| Trim and hardware | A zipper, webbing or buckle from another vendor because the first was late | The pull feels different, the webbing shade is off, and failures appear in month three | Name every component to a vendor part number and require a written change note |
| Print and embroidery | A new screen, a new thread lot, or a re-digitised file | Logo placement shifts and the stitch density looks different under retail light | Hold the screen and the digitising file on record and approve a strike-off every run |
| Foam and stiffener | A different density or another source of EVA or PE board | The bag stands differently and it protects less than the approved article did | Write density in kilograms per cubic metre and a thickness tolerance |
| Packing and folding | A different fold, polybag or carton because the packing crew changed | Creases that were never there before, and presentation complaints from the retail floor | Photograph the fold and keep the packing instruction inside the file |
The reference was never written
Most programmes have an approved sample and no written specification. The sample is an object that ages, gets used and gets lost; the specification is what survives it.
The dye lot was approved but not recorded
A shade band without a lot number tells you what the colour was and not how to make it again. The number is the recipe; the band is only the photo.
A component was substituted under time pressure
Substitution happens on the floor when a part is late. Without a written change note it never reaches the file, so run five inherits it as if it were approved.
The style moved to a different line
A second line means different machines, a different feed setup and a different crew. The first-off check exists precisely to catch this and is often skipped on a repeat.
Nobody kept a unit from the previous run
Comparing run five to run one is impossible if run one no longer exists. Retain two units from every run: one sealed, one that gets used.
The tolerance was never stated
Without a stated tolerance, every comparison is an argument about opinion. A number turns the argument into a measurement.
Why Does the Same Dye Recipe Produce a Different Colour?
Because a recipe is a set of conditions, not a formula. The same weigh-out in a different bath, at a different liquor ratio, with a different water supply or a different dye shipment, lands somewhere else on the colour solid. Dye houses control this by running a lab dip first and correcting it, which is why shade approval exists and why it takes the days it takes. Removing the dip does not make the recipe more stable; it removes the step that finds the instability.
The second reason is substrate. A bag is rarely one material in one colour. The shell, the webbing, the binding, the thread and sometimes the lining all have to sit together, and they are dyed in different places by different processes with different fibre contents. A polyester shell and a nylon webbing in the same visual colour are two different recipes. Matching them under one light source and calling it done is how a bag looks correct in the factory and wrong in a shop window, which is the problem metamerism causes and the reason a light box with more than one source is not optional.
The control is a recorded dye-lot number plus an approved shade band. The number is the recipe reference: it tells the dye house which bath made the approved fabric, and it is the only thing that lets run five be booked against run one instead of against a fresh guess. The band is the tolerance: the approved dip plus the limits either side of it that are still acceptable. Neither works alone, and a programme that approves colour without capturing both has approved nothing durable.
Does the Machine or the Operator Change the Finished Bag?
Yes, measurably, and in ways that show up as appearance complaints rather than as failures. Stitch length is set on a machine; seam allowance is set by a guide or by habit; topstitch alignment is set by a presser foot and by who is guiding the panel. Move a style from a single-needle lockstitch machine to a different class because the first one is busy, and the seam looks different even though it tests the same. That is not a defect by any written standard and it is exactly what a brand notices first.
The practical answer is not to demand that nothing ever changes, because lines do go down and crews do rotate. It is to require a first-off check at the start of every run, including reorders, and to hold the output of that check as a physical unit. A first-off is the first unit off the line after the machines are set; it is compared to the gold sample before the run continues. It costs an hour. It is the single highest-value control in the whole schedule and it is the one most often dropped on a repeat because the style is "known".
There is a naming convention worth insisting on. Write the stitch specification rather than describing the seam. A stitch type under ASTM D6193 with a stated stitch count per inch and a stated seam allowance is checkable by anyone with a ruler. "Stitched neatly" is not checkable by anyone, and it is what most tech packs still say in the one place that decides how the bag looks.
What Must a Spec File Contain Before the Second Run Is Booked?
A spec file is the written record of what was approved, complete enough that a stranger could rebuild the bag from it. An approved sample is not a spec file: it is one object, at one moment, with no tolerances and no supplier references. The test of a spec file is simple. Hand it to someone who has never seen the bag and ask them to quote run six. If they have to ask a question, the file is missing a field.
| Field | What it must say (worked example) | Why it decides run five |
|---|---|---|
| Shell fabric | 600D polyester, PU back-coated, piece-dyed to shade band A-2, dye-lot number recorded | A fabric code without a lot number cannot be re-booked, only re-guessed |
| Lining | 210D polyester, colour reference, mill code | The lining is usually the first thing substituted because nobody wrote it down |
| Webbing and binding | 25 mm polyester webbing, colour reference, vendor part number | Webbing is dyed elsewhere; without a part number it arrives a shade off |
| Foam and stiffener | 5 mm closed-cell EVA, 55 kilograms per cubic metre, plus or minus 0.3 mm | Density and thickness decide protection and stand-up, and both drift silently |
| Hardware | No. 5 nylon coil zipper, branded slider, vendor part number, finish stated | The vendor part number is what stops a similar slider being fitted in a hurry |
| Thread | Bonded polyester, ticket number, colour reference | Thread lot changes are visible on topstitch and invisible in the file |
| Stitch specification | ASTM D6193 type 301, eight stitches per inch, 10 mm seam allowance | A number can be re-set on a machine; an adjective cannot |
| Dimensions and tolerance | 330 x 240 x 180 mm, plus or minus 5 mm on each axis | Without a tolerance every dimensional comparison becomes an argument |
| Weight tolerance | Empty weight plus or minus 5 percent | Weight drift is the earliest signal that a material was substituted |
| Print and embroidery | Screen reference, colour reference, placement from a named datum, strike-off on file | A re-digitised file moves the logo and nobody notices until retail |
| Packing instruction | Fold method, polybag specification, carton marks, carton quantity | Presentation complaints are packing failures, and packing is rarely specified |
| Gold sample reference | Sample revision 2, sealed on a named date, one held by each party | Two identical sealed references remove the argument about which one is right |
| Change-control rule | No substitution without a written change note and a rebuilt pre-production sample | This is the clause that makes every other field enforceable |
The six fields nobody writes down
- The dye-lot number of the approved fabric, not just the fabric code and the shade name.
- The vendor part number for every piece of hardware, including the slider and the pull, not just the zipper size.
- The density of the foam and the stiffener in kilograms per cubic metre, with a thickness tolerance.
- The stitch type and the stitch count per inch at each named seam, referenced to a recognised stitch standard.
- The fold method and the polybag specification, because presentation complaints are packing failures.
- The change-control rule itself: no substitution without a written note and a rebuilt pre-production sample.
How Is a Gold Sample Used, and Who Holds It?
A gold sample is the sealed physical reference that every later run is judged against. It is not the sample you carried around the office and it is not the one on the shelf in the showroom. It is a unit from the approved run, sealed and dated, held in two places: one with the brand and one with the production team. Two copies exist so that a dispute is settled by comparing two objects rather than by arguing about one.
Three units per run is the working rule. One is sealed and never opened. One is used — carried, loaded, washed if the product is washable — because a used unit tells you what the field does to the bag and a sealed one only tells you what you approved. One is cut open, photographed and filed, because internal construction is invisible from outside and is the first thing to be quietly changed. Three units out of 500 is a rounding error on the order and it is the whole insurance policy on the programme.
Retention has a shelf life and it should be stated. Fabric yellows, elastic relaxes and adhesives migrate, so a gold sample from three years ago is a reference for construction and dimensions but a poor reference for shade. The practical rule is to refresh the sealed reference every two years from a current run, and to keep the original shade band and lot record permanently, because those are documents and they do not age.
What Is a Defensible Colour Tolerance, and How Is It Measured?
A tolerance is defensible when both sides can measure it with equipment they already own and when the number is big enough to be achievable and small enough to be invisible. "Match the sample" is not a tolerance. "Within the approved shade band, judged under a light box at a stated source, at an AATCC grey-scale rating of 4 or better" is a tolerance, and it can be settled in five minutes by two people who disagree.
| Attribute | Defensible tolerance on a repeat | How it is measured | What happens outside it |
|---|---|---|---|
| External dimensions | Plus or minus 5 mm on each axis | Tape on a unit laid flat and unloaded | Fit complaints, shelf presentation problems and retail rejection |
| Empty weight | Plus or minus 5 percent | Scale, same unit, no contents | Freight quotes drift and the spec sheet stops matching the product |
| Shade | Inside the approved band, grey-scale rating 4 or better | Light box with the band and the sealed reference alongside | Visible mismatch between cartons in the same shipment |
| Stitch count | Plus or minus one stitch per inch | Counted over 25 mm at the named seam | Seam appearance changes and seam strength moves with it |
| Foam thickness | Plus or minus 0.3 mm | Calliper on the cut-open reference unit | Protection drops and the bag stops standing the way it did |
| Hardware finish | No visible deviation from the retained first article | Visual comparison under the same light source | Brand perception damage on a bag that passed every written test |
| Print placement | Plus or minus 2 mm from the named datum | Ruler from the datum stated in the pack | Logo looks off-centre and every photograph has to be re-shot |
The grey scale deserves one clarification because it is where arguments start. The AATCC grey scale for colour change runs from 1 to 5, with 5 being no change, and a rating of 4 means a slight but perceptible difference. Accepting 4 or better on a commercial bag programme is realistic; demanding 5 is demanding a re-dye on almost every lot, and the cost of that lands in the unit price rather than in the colour.
How Do You Stop Silent Trim and Hardware Substitution?
Silent substitution is the highest-damage drift because the bag passes inspection and fails in the field. A zipper from a different vendor has the same size, the same colour and a different cycle life. A buckle from a different tool has the same shape and a different wall thickness. None of that is visible in an AQL 2.5 inspection, which samples for appearance and workmanship, and all of it shows up months later as a return.
Three controls work and none of them costs anything. First, name every component to a vendor part number in the spec file, including the parts people do not think of as components: the slider, the pull, the zipper tape, the thread. Second, require a written change note before any substitution, with the reason and the replacement part number on it. Third, hold a retained first article of every piece of hardware, so a substituted part can be identified by comparison rather than by argument.
The fourth control is the one that makes the other three bite: make the change note a condition of payment. If a substitution without a note is a breach with a contractual consequence, the decision stops being made on the floor at six in the evening and starts being made in writing. That is not an adversarial clause. It is the mechanism that lets a production team say no to a deadline without losing the order.
How Do You Write Change-Control Rules Into the Contract?
Change control is one clause with three parts: what counts as a change, what has to happen before the change is made, and what happens if it is made anyway. Most contracts have none of the three, which is why a substituted buckle is discovered by a customer rather than by a document. The clause is short and it belongs in the purchase order, not in an appendix nobody reads.
- Define a change. Any change to material, component, vendor, construction method, packing or artwork is a change. A change of vendor for an identical part is a change.
- Require written notice before the change, not after. Notice carries the reason, the replacement part number, and the effect on price and on the delivery date.
- Require a rebuilt pre-production sample for any material or component change. The new article is built from the new material and approved before bulk starts.
- Require a new first-off check for any construction or line change. A style moved to a different line is a construction change even when nothing else moved.
- Give the buyer the decision. A change notice that nobody has to answer is a change notice that will be treated as approval. State a response window and a default.
- Attach a consequence. An unnotified substitution is a breach with a stated remedy. Without a remedy the clause is a preference.
- Restart the clock on the date of the change. Any approved change restarts the bulk window from the date it is signed, in writing, with a new date.
The restart clause is the one that protects both sides. A buyer who approves a material change on day 20 of a 35-50 day window cannot reasonably expect the original date to survive, because a new material puts a fresh thread on the critical path. Putting that in writing at the start removes the argument before it exists, and it is the same clock rule that makes the 35-50 day schedule checkable in the first place.
What Does the Fifth-Reorder Verification Consist Of, Step by Step?
The fifth reorder is the test because by then every shortcut in the file has had four runs to surface. The verification is a fixed sequence of ten steps, run in the same order every time, producing one document. It is not a more thorough inspection; it is a comparison, and it can only be run if the gold sample and a retained unit from run four both still exist.
| Step | Action | Acceptance | Evidence produced |
|---|---|---|---|
| 1 | Lay out the gold sample, the retained run-four unit and one unit from run five | All three present and identified | Dated photograph of the three together |
| 2 | Compare the shell shade under a light box against the approved shade band | Inside the band, grey scale 4 or better | Recorded rating and a photograph under the stated source |
| 3 | Measure the three external dimensions and the strap length on all three units | Within plus or minus 5 mm | Measurement sheet with all three recorded |
| 4 | Weigh the empty units | Within plus or minus 5 percent | Scale readings recorded against the spec |
| 5 | Count the stitches per inch at the named seam | Matches the stitch specification | Close-up photograph with a scale in frame |
| 6 | Check every hardware item against the vendor part numbers in the file | No substitution, or a change note on file | Component checklist signed and dated |
| 7 | Open one unit and compare the internal construction to the cut-open reference | Foam, stiffener and seam construction match | Cut-open photograph beside the reference |
| 8 | Run the functional test named in the pack | Zipper cycles, load or drop protocol pass | Test record with the date and the result |
| 9 | Draw the AQL 2.5 sample and inspect the lot under ISO 2859-1 | Ac/Re decision recorded | Signed inspection report |
| 10 | Sign the release and file it against the dye-lot number | Release signed by the buyer | Release note carrying the lot number |
Step 10 is the one people skip and it is the one that makes run six work. Filing the release against the dye-lot number means that in eight months, when run six is booked, somebody can look up which bath made the approved fabric. Without it the programme restarts at step 2 with a guess, which is precisely the drift this whole page exists to prevent.
How Many Units Should You Pull for a Reorder Comparison?
The comparison sample and the AQL sample are two different things and they should not be confused. The comparison sample is small, fixed and judgemental: three units, chosen at random from the run, measured against the gold sample. The AQL sample is statistical: its size comes from the lot size under ISO 2859-1 at General Inspection Level II, and its purpose is to decide whether the lot as a whole passes.
Three units is the right comparison size for a practical reason rather than a statistical one. You need one to keep, one to cut open and one to photograph, and beyond three units you are duplicating work that the AQL sample already covers. What matters more than the count is that the three are pulled at random and pulled after packing, because a comparison run on hand-picked units from the top of the first carton tells you about the top of the first carton.
For the statistical side, the sample size is not a matter of preference. A lot of 500 pieces at General Inspection Level II draws a specific sample size with a specific accept and reject number at AQL 2.5, and the AQL sampling table gives the exact figure for any lot quantity. The point of drawing it properly is that the verdict means something: a lot passed at a stated sample size and Ac/Re pair is a defensible statement about the whole run, and a lot "looked at" is not.
What Happens When a Component Is Discontinued Mid-Programme?
It happens to every programme that runs long enough. A zipper vendor drops a finish, a webbing colour is retired, a mould wears out. The question is not whether it happens but what the procedure is when it does, and the procedure should be written before it is needed because a discontinuation always arrives with a deadline attached.
- Treat discontinuation as a change, not as a substitution. It goes through the same written change note with the same approval, even though nobody chose it.
- Require the vendor part number of the replacement and a first article. "Equivalent" is not a part number, and a first article is what makes the comparison possible next time.
- Rebuild the pre-production sample from the replacement. One unit, built from bulk material, approved in writing, filed with the new part number.
- Re-approve the shade if the component is visible. A new webbing lot in the same nominal colour is a new dye lot and it needs to be judged against the band.
- Restart the clock from the date the replacement is approved. The new component puts a procurement thread back on the critical path.
- Update the spec file in the same week. A replacement that exists only in email is a drift waiting for run six.
- Order a buffer if the replacement is also time-limited. Some programmes buy a two-run buffer of an irreplaceable component rather than re-engineering twice.
The last item is worth more attention than it usually gets. When a component is genuinely irreplaceable and the programme has years left, buying two runs of it at once is frequently the least expensive answer. Storage cost is small next to a re-engineering cycle, a new pre-production sample and a fresh round of approvals.
Does Consistency Cost More Than Accepting Drift?
Up front, yes, and not by much: a first-off check costs an hour per run, three retained units cost three units, a rebuilt pre-production sample on a material change costs the sample window inside the 35-50 day bulk range. None of those is a line item a buyer would notice on a 500-piece order. What costs money is drift, and the bill arrives in the three places that are hardest to attribute: returns, re-photography and a reorder that becomes a negotiation instead of a purchase order.
The comparison that matters is the cost of the fifth reorder with and without a spec file. With one, run five is a purchase order against a recorded dye lot, a known line and a stated tolerance; it clears in the normal 35-50 day window and it is inspected once. Without one, run five is a fresh development cycle: new dips, a new pre-production sample, a new argument about how run one actually looked, and a real chance the goods are rejected at inspection and re-worked. The second path costs several times the first and it delays the launch by more than the sample window ever would.
- MOQ
- 500 pcs per style / colourway (300 on a repeat)
- Sampling
- 6-10 working days (12-15 for complex builds)
- Bulk production
- 35-50 days from sample approval
- Repeat order reference
- Dye-lot number, gold sample revision and vendor part numbers on file
- Inspection
- AQL 2.5, ISO 2859-1, General Inspection Level II
- Stitch reference
- ASTM D6193 stitch type with a stated stitch count per inch
- Payment
- T/T 30/70
- Incoterm
- FOB Xiamen (CIF / DDP available)
- Operating since
- 2014
Who Signs Off a Reorder, and What Evidence Do They Need?
One named person, on the buyer side, with the authority to accept or reject and the discipline to sign. Reorder sign-off fails when it is diffuse: a production team releases the goods, a buyer sees them at the warehouse and a brand manager sees them at the launch. By then the only available remedy is a claim, which is the most expensive way to resolve anything. A single signer with a checklist converts a claim into a decision made before shipment.
The evidence set is short and it should be a physical folder rather than a thread in an inbox. It contains the spec file with the current revision date, the dye-lot record for the run just made, the component checklist with the vendor part numbers, the comparison measurements from the three-unit check, the first-off record, and the signed inspection report with the Ac/Re decision. Six documents. If one is missing, the honest answer is to hold the release rather than to release and hope.
Send run one back to us with the file
If you already have a programme and you are not certain what is written down, send us the approved sample, the current purchase order and any tech pack you have. We will come back with the spec file fields that are missing, the tolerances worth stating, and the change-control clause to put in the next order — sampling 6-10 working days, bulk 35-50 days from approval, inspected at AQL 2.5 under ISO 2859-1, quoted FOB Xiamen.
Frequently Asked Questions
Why does the fifth reorder look different from the approved sample?
Because the reference was never written down. By run five the fabric lot, the hardware vendor, the line and possibly the screen have all changed without a document. The approved sample is an object that aged; a spec file with a dye-lot number and vendor part numbers is what lets run five be built to run one.
What is a gold sample and how is it different from an approval sample?
An approval sample is what you signed off at the start. A gold sample is a unit from that approved run, sealed and dated, held in two places so a dispute is settled by comparing two objects. Refresh the sealed reference about every two years, because fabric yellows and adhesives migrate.
How many units should I keep from every production run?
Three. One sealed and never opened, one used in the field, and one cut open and photographed so the internal construction is on record. Out of 500 pieces that is negligible in cost and it is the only way run six can be compared to run five.
What causes colour to drift between reorders?
A dye recipe is a set of conditions, not a formula: a new bath, a different liquor ratio or a different dye shipment moves the result. Multi-component matches drift more, because the shell, webbing and binding are dyed in different places. Record the dye-lot number and keep an approved shade band.
What colour tolerance should I accept on a repeat order?
Inside the approved shade band, judged under a light box at a stated source, at an AATCC grey-scale rating of 4 or better. The scale runs 1 to 5 with 5 being no change; demanding 5 forces a re-dye on almost every lot and the cost lands in the unit price.
How do I stop a supplier substituting hardware without telling me?
Name every component to a vendor part number, require a written change note before any substitution, hold a retained first article of each part, and attach a contractual consequence to an unnotified substitution. The consequence is what moves the decision off the floor and into writing, where you can see it.
What should change control say in the purchase order?
Three things: what counts as a change, what must happen before it is made, and what happens if it is made anyway. Include a rebuilt pre-production sample for any material change, a new first-off check for any line change, and a restart of the clock from the date the change is signed.
Can I keep the same dye lot for every reorder?
You can re-book the same recipe by lot number, which is the point of recording it. You cannot keep the same physical bath from run to run. What the lot number gives you is a recipe the dye house can repeat, plus a reference to judge the result against.
Does a reorder still take 35-50 days?
Usually yes, and sometimes slightly less when no tooling or new screen is needed. The material thread still runs from the mill, dyeing still runs its course, and inspection still has to be drawn and reported. Tooling disappears from the schedule on a repeat, which is where the saving sits.
What is the minimum order quantity on a reorder?
500 pcs per style and colourway, and 300 is possible on a repeat because the set-ups already exist. Each colourway still carries its own dye lot and its own print or embroidery set-up, so the reduction reflects the absence of development work rather than a smaller run being easy.
How do I check a reorder without travelling to inspect it?
Ask for the ten-step comparison: three-unit layout against the gold sample, shade rating under a light box, dimension and weight measurements, stitch count, hardware part-number check, cut-open construction photograph, functional test, and the AQL 2.5 report. Every step produces a document, so the whole check runs from a folder.
What is a first-off check and why does it matter on a repeat?
A first-off is the first unit off the line after the machines are set, compared to the gold sample before the run continues. It catches a machine change or a crew change at hour one. It is skipped most often on repeats, which is exactly when a style has moved line.
What happens if a component is discontinued mid-programme?
It goes through the same written change note as any other change. Get the replacement part number and a first article, rebuild the pre-production sample from the replacement, re-approve the shade if the part is visible, and update the spec file in the same week.
Should I buy a buffer of an irreplaceable component?
Often yes. When a part is genuinely irreplaceable and the programme has years left, two runs of storage cost less than a re-engineering cycle plus a new pre-production sample plus a fresh round of approvals. Make the decision before the discontinuation notice arrives, because it always comes with a deadline.
How is the stitch specification written so it survives a reorder?
As a stitch type under a recognised stitch standard with a stated stitch count per inch and a stated seam allowance, at each named seam. An adjective such as neatly cannot be re-set on a machine; a number can, which is why it survives a change of line or operator.
Does a substituted component show up in an AQL 2.5 inspection?
Usually not. An attribute inspection samples for appearance and workmanship, and a substituted zipper or buckle looks correct and measures correct, so it passes. That is why the control sits in the component checklist and the written change note rather than in the inspection report.
Who should sign the release on a reorder?
One named person on the buyer side, with authority to accept or reject, working from a six-document folder: the spec file, the dye-lot record, the component checklist, the comparison measurements, the first-off record and the signed inspection report. If one is missing, hold the release.
Can a reorder be inspected at a different AQL level than the first run?
It can, but it should not be changed without a reason on the record. Changing the level changes the sample size and the accept and reject numbers, which means the two runs were not judged by the same rule. Keep AQL 2.5 under ISO 2859-1 at General Inspection Level II unless there is a documented reason.
People Also Ask
Why do reorders look different from the original sample?
Because eight variables move between runs: dye lot, fabric batch, machine, operator, trim and hardware, print and embroidery, foam and stiffener, and packing. Each one needs its own written control, or it drifts silently.
What should a bag spec file contain?
Shell and lining with mill codes, webbing and binding part numbers, foam density and thickness tolerance, hardware part numbers, thread, stitch type and count, dimensions with tolerance, weight tolerance, print references, packing instruction, gold sample reference and the change-control rule.
What is a gold sample in bag production?
A sealed, dated unit from the approved run, held in two places, that every later run is judged against. Keep three units per run: one sealed, one used in the field, one cut open and photographed.
How do you verify the fifth reorder?
Ten steps: lay out gold sample, retained prior unit and new unit; rate shade; measure dimensions; weigh; count stitches; check hardware part numbers; cut open one unit; run the functional test; draw the AQL 2.5 sample; sign the release and file it against the dye-lot number.
How do you stop hardware substitution on repeat orders?
Name every component to a vendor part number, keep a retained first article of each, require a written change note before any substitution, and attach a contractual consequence to an unnotified substitution so the decision is made in writing rather than on the floor.
Sources and further reading
Every figure on this page is either our own programme data or attributed below. Open the originals before you write them into a specification.
- ISO 2859-1:1999 Sampling procedures for inspection by attributesInternational Organization for Standardization
- ASTM D6193 — Standard Practice for Stitches and SeamsASTM International
- AATCC technical manual: grey scale for colour change and colour assessment proceduresAmerican Association of Textile Chemists and Colorists (AATCC)
- ISO 9001:2015 Quality management systems — Requirements (documented information and change control)International Organization for Standardization