Home › Field notes › YKK vs Generic Zipper Sourcing: Mating Life, Tooth Pitch, Salt Spray

Modular Bag Pro · 2026-10-10 · 3,056 words · 14 min read

A branded zipper chain and a generic chain differ mainly in tolerance control and traceability: an established supplier in that family holds tooth pitch, slider mouth geometry and plating thickness inside published limits and can produce lot records, while an unbranded chain may meet the same nominal size with wider variation and no paperwork. Four measurements separate them in a buying decision - mating cycles before the chain stops closing cleanly, pitch drift along a 600 mm length, hours to first base-metal corrosion under ASTM B117 neutral salt spray, and the width and pick count of the tape. Hardware programmes run at MOQ 500 per reference, sampling across 6-10 working days and bulk over 35-50 days, released at AQL 2.5. Nothing here is a claim of agency for, or use of, any zipper brand; the scope is civilian carry - travel, commuting, worksite and first-aid equipment.

What a chain, a tape, a slider and a puller each contribute

A zipper is four parts working as one mechanism, and a specification that names only the chain size has specified roughly a quarter of the item. The chain carries the closing load. The tape transfers that load into the seam. The slider does the actual work of interlocking and separating the teeth, and its mouth geometry determines whether the chain feeds smoothly or jams. The puller is the human interface and, in practice, the part a customer touches first and complains about first.

An established supplier in the branded family - the group that includes YKK and a small number of comparable industrial names - controls all four parts as one system and publishes the tolerances it works to. Tooth pitch is held on tooling that is measured and maintained, tape is woven to a stated width and pick count, slider bodies are die-cast or formed on dedicated tools with a defined mouth profile, and finishing runs to a stated plating thickness. The generic route buys the same nominal size from whichever mill has capacity, which is not the same thing as buying the same part.

The consequence for a buyer is that a generic chain can be perfectly acceptable and still be the wrong choice, because acceptability is a per-lot property when it is not underwritten by a controlled process. A branded chain is boring in the best sense: the tenth thousand behaves like the first hundred. Where the product is sold with a warranty, or where a failure means the whole bag comes back rather than one component being replaced, that boring predictability is worth more than the unit saving.

Verdict: Treat the four parts as one controlled assembly and specify all of them - chain size, tape width and pick count, slider type and mouth profile, puller finish - because a saving taken on any single part is paid back through the seam, the slider or the returns desk.

Mating life: how many cycles before a chain stops closing cleanly

Cycle life is the number that most buying decisions claim to rest on and the one most often quoted without a method. A cycle count means nothing until the stroke length, the load on the chain, the operating speed and the failure criterion are all stated. A workable internal method runs a 300 mm stroke at a controlled rate with a 2 kg transverse load across the closed chain and counts cycles until the slider no longer closes the chain fully on a single pass, or until tooth wear becomes visible at the point where the slider reverses direction.

On that kind of rig a well-made moulded-tooth chain in #5 usually reaches several thousand cycles before the criterion is met, and a metal chain of the same size frequently runs longer; a poorly controlled chain can start skipping inside the first thousand. The relevant point for a specification is not the absolute figure but the gap between the best and worst sample in a lot, because the gap is what the customer experiences. A lot whose worst sample fails at 40 per cent of its best sample is a lot with a field-failure tail, and no amount of average performance removes that tail.

Two practical notes follow. First, cycle testing belongs on the finished bag, not only on a loose chain, because the tape is stitched under tension in the product and a chain that cycles well on the bench can bind once the seam pulls it into a curve. Second, the failure that ends a product's life in the field is rarely the teeth; it is a slider that has lost its crimp, a puller that has bent, or a bottom stop that has walked off the tape.

Spec rule: State the cycle method in the specification - stroke length, transverse load, rate, sample count per lot and the exact failure criterion - and accept a chain only when the worst sample in the lot, not the average, meets the stated count.

Tooth pitch consistency and why a drifting pitch jams a slider

Tooth pitch is the distance from one tooth to the next along the chain, and it is the single dimension most responsible for smooth feeding. Measure it at ten points along a 600 mm length under magnification and the difference between the two routes becomes obvious: a controlled chain holds pitch within a tight band along the whole length, while a chain made on worn or shared tooling drifts, usually widening toward the end of a run as the forming tool heats.

The slider is a fixed-geometry part. Its mouth was made for one pitch, and it tolerates variation only up to the point where two teeth try to enter the mouth at once or a gap lets a tooth arrive out of sequence. The symptom the user reports is not "pitch drift" but a zipper that catches halfway, needs a second attempt, and eventually splits behind the slider - which is the failure mode that generates returns, because the bag appears to be closed and then opens.

Pitch also interacts with curvature. A chain sewn around a corner is asked to open slightly on the outside of the curve and close on the inside; a chain with consistent pitch does this evenly, and a chain with drifting pitch binds at the point where the drift and the curve coincide. Product designers notice this as "the zipper catches at the same place every time", which is a geometry problem rather than a hardware defect, and it is solved either by tightening the tolerance or by moving the radius.

Takeaway: Measure tooth pitch at ten points over a 600 mm length and reject any lot whose spread exceeds the stated band, because pitch drift is the root cause of the catching and splitting failures that customers describe as a broken zipper.

Salt spray behaviour: what ASTM B117 proves and what it does not

Corrosion testing is where hardware comparisons go wrong most often, usually because a result is read as a lifetime prediction when it is a comparative screening result. Neutral salt spray to ASTM B117 exposes parts to a continuous fog at a controlled temperature and reports the appearance of corrosion after an agreed number of hours; the equivalent international method is ISO 9227. What the result tells you is how two finishes compare under one aggressive, constant condition. It does not tell you how many years a part lasts on a bag carried in a coastal city, because real exposure alternates wet and dry, includes ultraviolet light and abrasion, and never holds a constant temperature.

Used correctly, the method is still the most practical way to compare plating routes. A bright nickel finish on a brass or zinc-alloy tooth will usually pass a short exposure and then show spotting; a black-oxide or epoxy-coated finish behaves differently again, often holding appearance longer but failing by coating breach rather than by uniform corrosion. The failure that matters for a bag is the one that marks clothing or hands, which is why appearance criteria and a rub test after exposure are more useful than a bare hours figure.

The specification clause should therefore name the finish, the exposure hours, the acceptance appearance and a post-exposure handling check, and it should state explicitly that the result is comparative. Two further controls matter more than hours: the plating vendor must be named and held constant, and any change of vendor triggers a re-test, because the same nominal finish from two vendors rarely performs alike.

Judgement: Write the salt spray clause as a comparative screen - finish named, exposure hours stated, acceptance appearance defined, plating vendor named and any vendor change triggering a fresh test - and never present the hours figure to a customer as a service-life prediction.

Tape, slider body and puller: where the unit cost is really saved

When a generic chain undercuts a branded one, the saving is taken somewhere specific, and knowing where lets a buyer decide whether the saving is acceptable. Tape is the most common place. A narrower tape or a looser pick count costs less per metre and still sews, but it holds stitch less well, frays at the bottom stop earlier and lets the chain roll under load. The second common place is the slider body: a thinner die-casting, a less precise mouth and a weaker crimp all reduce cost and all reduce cycle life. The third is the puller, where a lighter gauge and a thinner plating are almost invisible on the shelf and very visible after a season.

There is also a legitimate saving that has nothing to do with quality: packaging, freight consolidation and the absence of brand licensing cost. A generic chain bought in one size and one colour in a single large lot can be genuinely cheaper for reasons of logistics alone, and for a product where the zipper is not a selling point that saving may be the right decision. The discipline is to identify which kind of saving is on the table before accepting it.

Chain type matters as much as brand. Moulded polymer teeth in #5 are quiet, light and free of any corrosion mechanism, which suits travel and commuter products. Metal teeth in the same size suit work and heritage styling but add mass and need a finish clause. Coil chains - the spiral construction often used on curved openings - feed around corners better than either and are the usual answer where an opening wraps a panel. Selecting on brand before selecting on chain type gets the decision backwards.

Bottom line: Identify where the saving comes from - tape width and pick count, slider casting and crimp, or puller gauge and plating - and accept a generic chain only when the saving is logistical rather than structural, and when the chain type itself suits the opening geometry.

Sourcing and lead-time risk: allocation, colour minimums, dual sourcing

Supply risk, not quality, is usually the reason a programme ends up dual-sourcing hardware. Established zipper suppliers run allocation in peak months: when the cut-and-sew season fills, popular sizes and colours are rationed to contracted customers first, and a small programme without a booking can wait. The generic market behaves differently but not better - capacity is available, yet the specification drifts between lots, so the risk arrives as variation rather than as delay.

Colour is the second constraint. A branded chain in a standard shade may be available from stock in a low minimum, while the same chain in a matched shade carries a dye-lot minimum that can exceed the bag MOQ of 500 units by a wide margin. Practical responses are to accept the nearest standard shade from the supplier's card rather than matching the shell exactly, to design the product so the tape is concealed, or to pool one tape shade across several references so the dye-lot minimum is met once.

Dual sourcing is the standard answer and it needs designing in, not improvising. Keep two approved suppliers on the bill of materials for every chain size used, hold a physical reference sample from each, and record the differences - tape shade, slider marking, puller profile - so that a substitution is visible on the line rather than in the field. Where a puller carries a moulded logo, dual sourcing means two tools, and the second tool is cheap insurance against a line stoppage.

Selection rule: Keep two approved suppliers on every chain size, book hardware when the sample is released rather than after approval, choose tape shades from the stock card where the dye-lot minimum would otherwise exceed MOQ 500, and hold a physical reference from each supplier on the line.

How to write a zipper clause a second supplier can quote against

A hardware clause that names a brand and nothing else cannot be quoted by anyone else, and a clause that names only a size can be met by almost anything. The workable form is brand-or-equal: name a reference supplier as the benchmark, then list the properties a substitute has to demonstrate, and state which documents prove it. The list normally covers chain type and size, tooth material, tape width and pick count, tape shade reference, slider type and lock function, puller style and finish, plating specification with exposure hours, cycle count with method, and the inspection level applied to incoming goods.

Revision identity closes the clause. Every hardware line on the bill of materials bears a suffix which moves whenever the supplier, the plating vendor, the tool set or the tape origin moves, and goods arriving under a new suffix are handled as a fresh part needing approval rather than as the old one continuing. Without that discipline a second-year reorder can be compliant on paper and different in the hand, which is precisely the sort of failure a returns analysis cannot explain.

Our production team coordinates hardware programmes through a verified 4,950 m² base where 149 machines are tended by 137 people across seven lines, with output quoted at 200,000 pieces monthly; the founder has worked in bag production since 2004 and the company dates from 2014. Incoming hardware is checked at AQL 2.5 under ISO 2859-1 level II with a Critical 0 / Major 2.5 / Minor 4.0 split, backed by a retained reference sample per lot. Platform options sit under modular travel backpack and tactical backpacks, and full programme routes are described on custom modular backpacks.

Write the zipper clause as brand-or-equal: a named benchmark supplier, a property list any substitute has to demonstrate, the paperwork that proves each property, and a suffix that moves whenever the supplier, plating vendor, tool set or tape origin changes.

Commercial mechanics for a hardware-led programme

Hardware sets the critical path more often than sewing does, so the calendar has to be built around it. Standard sizes and stock shades can be bought while the sample is still in work; matched shades, custom pullers and special plating run longer and should be ordered at sample release. Sampling then takes 6-10 working days for a conventional build, extending to 12-15 where a custom puller tooling or a new plating is involved, and bulk follows across 35-50 days largely determined by the slowest bought-in component.

Tooling is the cost item unique to hardware. A custom puller with a moulded logo, a shaped slider or a branded top stop falls into the USD 300-2,500 band depending on cavity count and material, and the tool is retained for the account across repeat seasons. A development charge of USD 50-150 per reference comes back to the buyer once the volume order follows. Every reference opens at 500 pieces minimum, priced on an indicative FOB Xiamen basis settled 30 per cent on deposit and 70 per cent before loading, with a quotation returned within 24-48 hours of a complete hardware specification.

Freight then closes the plan. By sea the transit runs 25-35 days and suits a first fill; by air it is 5-8 days and suits a launch date already committed; by express courier it is 3-5 days and suits approval sets or a single carton of sliders when an assembly line is waiting. Hardware is dense and small, so it is the one component where paying for air often makes sense while the finished bodies still travel by water, and planning that split deliberately avoids the familiar scene of sewn shells stacked and waiting for a slider.

Place the hardware booking as the sample goes out rather than waiting for approval, treat a custom puller tool in the USD 300-2,500 band as the cost of keeping a marked part dual-sourced, and send hardware by air at 5-8 days whenever finished bodies are already planned for sea transit at 25-35 days.