Custom Pet Carrier Safety Tethers

Pet carrier production desk · Updated 2026-10-07 · 14 min read

A custom dog carrier brief starts with the weight band, not the colour: strap width, base board thickness and mesh ratio all follow from how heavy the dog actually is.

What a tether is for, and what it is not

A safety tether exists to cover a gap in time, not a gap in structure. The gap is the period during which the closure is open and the animal is still inside or half inside — which, measured with real handlers rather than in theory, is somewhere between three and fifteen seconds every single time the product is used.

Everything a customer does with a carrier passes through that window. Opening the top at the vet, unzipping an end panel in a car park, cracking the door a hand's width at arrivals, reaching in to lift a nervous cat out — all of them are moments when the animal could leave and no closure is doing anything about it. The tether is the only line of defence in those seconds, and almost every escape story we hear from brands involves that window rather than a closure failure.

What a tether is not is a restraint during transport. It should never be relied on to hold an animal during movement, in a vehicle, or anywhere force may be applied suddenly from outside. Restraint during motion belongs to a crash-tested harness system, which is a different product with different testing, and conflating the two is how brands end up making a claim they cannot support.

Nor is it a substitute for a working closure, and this matters for copy as much as for engineering. A product described as secure because it has a tether invites owners to travel with the door slightly open, and the tether will not save them from every version of that behaviour. We advise clients to position the tether as a second line rather than a primary one.

Finally, a tether is not a lead. It is too short, it is anchored to the wrong thing, and using it to walk an animal places lateral load on components that were never rated for it.

The three moments that actually produce escapes

Field data from escort incidents clusters into three scenarios, and the differences matter because they ask different things of the tether.

The first is voluntary exit at the destination — the animal walks out before the handler is ready. Here what matters is whether the tether remains attached at both ends, and failures are overwhelmingly user-error: never clipped, or clipped to something that came loose.

The second is startled exit, typically triggered by noise in a terminal or clinic. This produces genuine dynamic load, and it is the case the strength calculation below is designed around.

The third is the accidental open — a zip worked partway by an animal, or a closure not fully engaged. Here the tether may see sustained rather than shock load, which favours a webbing with controlled elongation over one with none at all.

Length is the specification everyone gets wrong

Length looks like the least interesting number on a tether and it is the one that decides whether the tether works. It can be derived rather than guessed, which is what makes it worth doing properly.

The upper bound is set by geometry. The tether must not allow the animal's head and shoulders through the aperture it is guarding, because once the shoulders are out, the tether is holding nothing useful. Take a typical end aperture 250 mm tall and 300 mm wide, with the tether anchored centrally on the far seam. Trigonometry gives the ceiling: if the anchor sits 400 mm from the aperture plane, a tether length beyond roughly 260 mm permits the shoulder line to reach the opening on a medium dog.

The practical figure we work to is therefore 150 to 220 mm of usable length measured from anchor bearing surface to clip bearing surface, not overall webbing length before assembly. Anything above about 250 mm stops preventing exit and starts merely documenting it.

The lower bound is set by welfare rather than geometry, and it is tighter than most people assume. Too short and the animal cannot turn around, cannot lie in its preferred position, and — the serious one — experiences constant neck pressure whenever it shifts. Our minimum is the animal's own shoulder-to-shoulder width plus 60 mm of freedom, which for a mid-size dog lands near 170 mm and for a large cat near 140 mm.

That window between about 150 and 220 mm narrows fast at both ends of a size ladder. This is why we size tethers per size rather than specifying one length across a range, even though one length is cheaper and easier to buy. A tether fitted to the smallest size in a range is a welfare problem; one fitted to the largest is a safety problem.

Measuring usable length rather than webbing length

Most disputes about tether length turn out to be disputes about where the measurement starts. A procurement drawing may specify 200 mm of webbing; the assembled product delivers something rather different once the end hide is wrapped round a bar, stitched, bar-tacked, and then connected through a swivel and a clip.

Each stage consumes length. A 180-degree wrap round a 20 mm bar takes roughly 32 mm of webbing and yields no usable reach at all; the swivel body and its connection take a further 20 to 25 mm; the clip's own gate-to-bearing distance adds another 15 to 25 mm depending on the gate geometry. Specified as webbing length, a nominal 200 mm component routinely delivers 140 mm of usable reach, which is below the welfare floor for a medium dog.

Specify usable assembled length instead, and draw it on the sample report from bearing surface to bearing surface with the product in its shipping configuration. It is one more line on the drawing and it removes the single most common source of disagreement between a brand's expectation and what arrives.

Breaking strength, safety factor, and the dynamic spike

The number a hardware catalogue gives is a static break figure, and it is not the number to specify against. What a tether sees in service is a shock load, and the difference between the two is roughly threefold.

An animal launching from a standstill generates force well above its body weight. A 12 kg dog produces about 118 N standing still. Launching against a restriction, briefly and without build-up, measured peaks run to two and a half to three times body weight — call it 300 to 355 N for that animal. Add a handler's opposing pull and the peak approaches 425 N.

Against that, a webbing assembly refuses no meaningful safety factor below about 1,000 N. Our standard specification is 1,200 N minimum for mid-size formats and 1,500 N for large, giving a safety factor between 2.8 and 3.5 on the realistic peak rather than the more comfortable-looking and meaningless multiple of static weight.

Then there is the question most specifications omit entirely: where is the weak link? A tether can fail at the webbing, at the stitch, at the clip, or at the anchor in the bag, and every one of those except the last can be designed deliberately. We prefer the webbing to be the strongest element and the stitch the engineered control, because a progressive stitch failure gives warning whereas a clip fracture does not.

Elongation also deserves a specification, which few tethers carry. Some give is useful — it absorbs the shock peak and reduces what the anchor must resist — but too much turns a 180 mm tether into a 240 mm one precisely when it matters. We hold elongation to between 6 and 10 percent at 500 N, verified on a tensile rig rather than accepted from a datasheet.

Why the clip should be weaker than nothing at all

That heading is deliberately provocative and the point behind it is important: there is no safe failure mode for this component. A tether that releases under load produces the outcome it exists to prevent. So every element — webbing, stitch, clip, anchor — should be specified to the same safety factor rather than to its own convenience, and the assembly should be tested as an assembly.

In practice this means buying the clip to the same rating as the webbing rather than accepting whatever came attached to it, and it routinely doubles unit cost for that one component. We have never had a client regret it after seeing a jerk-test video.

Where the anchor goes inside the shell

Anchor position governs what the tether can physically achieve, and it is chosen last far more often than it should be. Three rules cover the decision.

Anchor behind the animal's shoulder line, not beside it. An anchor level with the animal's shoulders lets the animal pivot and use the full aperture arc. An anchor set behind the shoulder restricts forward travel much more effectively at the same tether length — typically removing 60 to 90 mm of effective reach, which is why it earns its place.

Anchor high, not at floor level. A floor-level anchor allows circling, and a circling animal winds a tether down to a fraction of its usable length, producing the neck pressure described above. An anchor at or above the animal's mid-body height keeps the tether clear of limbs.

Anchor to structure, not to lining. This is where the interface with the bag itself becomes the subject, and the anchor options are set out properly in our writing on collar attachment points. Summarised here: an anchor that loads a single layer of lining will pull through at figures far below the webbing rating, so it must reach a seam, a reinforcement band, or the baseboard.

The fourth consideration is count. Two anchor points — one each side of the aperture — allow the handler to keep the animal off-centre during entry and exit, which is genuinely useful in practice. They also double the number of things a user can attach incorrectly, so if two are specified they should be visibly keyed different.

Clips, swivels, and the deliberate-release question

The clip is the only part of the system the handler touches, and it has to reconcile two opposite requirements: it must not release under shock, and it must release when an adult needs it to under load.

A panic scenario makes the requirement concrete. An animal entangled, or a tether caught on something, and an adult needs to free it in seconds with one hand while managing the animal. Anything requiring two hands, fine motor control, or line-of-sight placement fails that test — and if it fails, the alternative an owner reaches for is a knife, applied to the product.

The components that satisfy both constraints are a bolt snap or a locking trigger snap in metal, sized so the release surface is large enough to operate under load. What does not satisfy them: a carabiner-style twist lock, which is slow and awkward one-handed; a plastic side-release buckle, which can release under a tangential shock; and any small-gauge clip where the release is a button under 8 mm.

A swivel is worth specifying and rarely is. Without one, the tether twists as the animal circles, and a twisted webbing loses effective load capacity while also shortening itself. A barrel swivel between anchor and webbing solves both for cents, and the argument against it — another component, another thing to buy — is weaker than the argument for it.

Finally: never attach to a collar. If an animal is wearing a collar and is tethered to it, a launch places the entire load across the trachea. The correct interface is a harness, and brands building a carrier programme alongside walking gear should read our note on custom harness development for how the two should be coordinated.

Chew resistance and why wire cable is the wrong answer

Animals chew tethers. This is not abuse, it is behaviour, and any specification that treats it as misuse has already lost. The design question is what happens when they do.

The tempting answer is steel wire cable sheathed in plastic, and it is the wrong one. Once an animal breaches the sheath — which takes minutes rather than months — the exposed strands form fine needles at the cut end. They lacerate mouths and tongues, they are invisible under casual inspection, and they produce an injury that reads as a product defect because it is one. We decline to specify wire cable in any pet carrier application, at any price point.

The better answer is flat webbing, which frays rather than spears, combined with either a chew-resistant sleeve or — preferably — a design that keeps the tether out of reach of the mouth when the animal is settled. The second is achieved by anchoring above the shoulder line and keeping the tether short, which together put it beyond comfortable reach for most animals most of the time.

Where additional protection is genuinely needed, the options are a tightly woven aramid-blend sleeve or a smooth TPU-coated webbing with no exposed edge. Both cost significantly more than standard polyester and both are appropriate for a product marketed for chewers. Neither removes the need for instruction: the tether should be described in packaging as a supervised-contact restraint, not as something to be left accessible indefinitely.

Nothing above should be read as permitting indefinite access. Whatever the construction, the tether belongs clipped during travel and released only when the handler has both hands free, and the care card should say exactly that in one sentence.

Safety tether specification table

The five configurations below cover the size ladder most ranges require. Usable length is measured assembled, bearing surface to bearing surface; break strength is assembly figure rather than webbing figure; elongation is measured at 500 N.

The medium-dog row is the one most programmes need and the one whose length — 190 mm — is most often specified wrong by brands reusing a single component across their ladder. Note also that the 15 mm row for cats reaches the same safety factor at lower strength because the animal generating the load is smaller; specifying the dog component on the cat product adds bulk to the animal's environment for no benefit.

Elongation deserves a second look in that table. The chew-resistant sleeve sets the lowest figure, which means it transmits the highest shock into the anchor — another reason its anchor interface needs to be specified alongside rather than inherited.

Whichever row applies, the figures only hold if the tether is actually attached. Write the advice on the care card as carefully as the component itself: a restraint nobody clips is a restraint with no rating at all.

Testing the assembly and releasing the shipment

A tether should be tested as bought, assembled to the bag it ships in, because every supplier's datasheet describes a component rather than a product. Four tests cover the real failure modes, and none of them is expensive.

Jerk test. Drop-weight rig producing a shock equivalent to three times animal body weight, ten repetitions, then inspect at 10x. Acceptance: no stitch movement, no clip deformation, no measurable permanent elongation beyond the specified figure.

Sustained load. 500 N held for five minutes, repeated hourly for twenty-four hours. This catches creep in plastic components and progressive stitch slip, neither of which appears in a snap test.

Chew simulation. A jaw fixture at typical bite force for the target weight class, fifty cycles against the mid-span and twenty against each termination. Acceptance is no breach to a load-bearing element, and anything that produces a sharp edge automatically fails.

Release test. Under 200 N of applied load, a single adult must be able to release one-handed in under five seconds, using the hand they would have free in reality.

At release we inspect to AQL 2.5 general inspection level II, and every tether assembly is checked rather than sampled regardless of the level, because it is a safety-critical component and inspection cost is trivial next to exposure. Critical defects are any clip that opens under a 150 N tangential pull, any stitch slippage visible without magnification, any missing swivel where specified, any anchor that moves more than 2 mm under 300 N, and any exposed edge on a chew-resistant sleeve.

Animal-welfare guidance belongs in the same conversation, and the frameworks our own protocols align with are those maintained by the Center for Pet Safety for restraint testing and the general product-safety expectations published by the U.S. Consumer Product Safety Commission. A brand quoting either to a retailer should be able to produce the test report rather than the component datasheet.

Terms follow the standard frame: MOQ 500 per colourway with the tether specified per size, samples in 6-10 working days, bulk production 35-50 days after approval. All coordination runs through an SGS-verified production base audited to BSCI and ISO 9001, with 7 lines and 137 people delivering 200,000 pieces monthly. Payment is T/T, 30 percent deposit and 70 percent before shipment, quoted FOB Xiamen.