Home › Field notes › Fidlock Technology Explained: How Magnetic Snap-Fit Works And Fails

Modular Bag Pro · 2026-10-10 · 2,754 words · 13 min read

Magnetic snap-fit works in two stages, and confusing them is the source of most bad specifications. First a magnetic field pulls the two halves into approximate register; then a mechanical undercut engages and carries the load once the halves are pushed, slid or tipped home. Fidlock technology is best known in bag hardware as one manufacturer of this family. The practical consequence for a designer is that the magnet never carries load, so the specification has to cover two separate things: enough field to locate reliably through the panel, and enough undercut geometry to hold the load case. Build it on 25 mm webbing with a controlled stack, and allow MOQ 500 per variant. Confusing the two stages is the source of most bad specifications: magnetic strength gets quoted where engagement force was needed, and the panel is never given a tolerance to hold. The nine sections below separate the mechanism from the marketing.

Stage one and stage two: guidance, then interlock

Stage one is magnetic. Two opposed poles, or a pole facing a ferrous element, generate an attractive field that acts over a short range. As the halves approach, the field produces a centring force: the halves are pulled toward the position where the field energy is lowest, which is roughly the aligned position. The user feels this as a light pull and a sense that the parts want to find each other.

Stage one is not retention. The centring force is small, falls off steeply with distance, and is easily overcome by a modest load or by a fabric fold pushing the halves apart. Its job is to remove the aiming step so that stage two can begin from a controlled starting position.

Stage two is mechanical. An undercut, hook, slide rail or cam captures the moving half against the fixed half and transfers load through polymer or metal rather than through the magnet. This is the part that matters for strength, for cycle life and for safety, and it is the part a specification should describe in numbers. Where a supplier's literature dwells on magnetic strength and says little about the interlock, treat that as a warning sign rather than as a feature.

The two-stage model also settles the argument about what a buyer may claim. "Holds strongly" is a statement about stage two and needs a force figure behind it. "Closes itself" is a statement about stage one and needs a register tolerance behind it. Products that advertise the first and are returned for failing the second are usually the victims of a specification that never separated them.

What a bag designer needs to know about the magnet

Three magnet properties matter in soft goods, and none of them is the headline strength figure. The first is grade, which sets how much field the material produces for its volume and how it behaves as temperature rises. Every magnet loses field as it heats, and some grades lose it sharply near their maximum operating temperature, so the cold-weather product that works well in a showroom may behave differently in a hot vehicle.

The second is field shape. A designer wants the field concentrated where the halves meet, not spread across the panel, because a spread field attracts loose metal items and gives the user a vague rather than a decisive pull. Shielding on the back face of the magnet is what keeps the field where it is useful, and it is worth asking whether a candidate part is shielded.

The third is what the field does to things around it. Modern chip cards and phone payment modules are not damaged by fields of the magnitude used in bag hardware, but magnetic stripe cards, mechanical compasses and some medical devices are affected by sustained proximity. Where a product will be sold into a channel that asks, state the position in the documentation: keep the fitting away from items the user may carry for medical reasons, and do not claim immunity you have not tested.

There is also a manufacturing handling issue that catches new programmes out. Loose magnets and magnetised parts behave badly on a cutting table and on a needle: they attract swarf, they cling to machine plates and they can deflect a needle if one is inadvertently introduced into a seam. Keep magnetised components in closed trays until the assembly step, keep them away from cutting and sewing stations, and count them in and out so a stray part is noticed rather than sewn in.

Selection rule: Treat the magnet as guidance hardware rather than as retention hardware, and size neither the panel nor the load until that separation appears in both the datasheet request and the technical pack.

Undercut geometry: how the load is actually carried

The interlock is where families differ, and the geometry dictates both the motion the user performs and the loads the fitting survives. Four motions cover most of the market, and each has a characteristic tolerance appetite and release behaviour.

Read the release column as a safety statement. A fitting that releases on a straight pull is the wrong choice for a pouch that will be grabbed and dragged; a slide or rotate motion resists accidental release far better because the releasing motion is not the motion a snag produces. Where a programme has had accidental-opening complaints, changing the release motion usually fixes it more effectively than increasing force.

Cycle life belongs to this geometry too. Wear occurs at the contact faces of the undercut, and the wear rate depends on the material pair and on how much sliding happens during engagement. Ask what the candidate's cycle figure was measured against and whether the failure at end of life was wear, cracking or loss of engagement force.

Contamination is the second variable in wear. Grit carried into the engaging faces acts as an abrasive paste and accelerates the loss of engagement force far faster than clean cycling does, which is why a laboratory cycle figure rarely matches field experience. Where the product will be used on trails, beaches or worksites, add a contaminated cycle sequence to the evaluation: cycle with a controlled amount of fine abrasive introduced, then re-measure.

Verdict: Undercut geometry carries the load while marketing rarely describes geometry precisely, so dimension the engaging faces rather than relying on a quoted pull figure borrowed from another panel build.

Tolerance stack-up: the conditions for reliable self-location

Self-location only works inside an envelope, and the envelope is set by the sum of several small tolerances. Panel thickness varies with material and with seam allowance. Stiffener thickness varies with the foam lot. Stitch tension varies with the operator and the machine. Housing seating depth varies with how hard the assembly step presses it. Individually each is under a millimetre; together they decide whether the halves meet.

Write the envelope as three numbers in the tech pack: the permitted total stack thickness under the flange, the permitted lateral offset between the two halves, and the permitted rotation of each half. Those three lines convert a hopeful assembly into a controlled one, and they give the production team something to check rather than something to feel.

Then control the inputs that feed the stack. Specify the stiffener as a material with a stated thickness and tolerance rather than as a description, cut the panel to a stated tolerance, and jig the housing position during assembly. Where two factories may build the product, run the stack check at both, because the same drawing produces different stacks on different machines.

Measure rather than trust. Take the stack reading on five production samples per shift during the first week, using a gauge rather than callipers where the material compresses, and plot the result. A spread that sits inside the envelope on Monday and drifts outside it by Friday is a process signal, and it is far cheaper to correct in week one than after a container has shipped.

Takeaway: Budget manufacturing variation explicitly by summing tolerances rather than assuming nominal parts, because self-location that succeeds at nominal dimensions fails once stack-up accumulates across a MOQ 500 run.

Orientation and handedness in a compliant panel

A chassis is not a rigid machine frame. It twists when worn, it bends when loaded and it relaxes over time. That compliance means orientation on the bench is not orientation on the body, and a fitting that self-locates perfectly on a flat table may behave differently when the panel is wrapped round a torso.

Handedness follows from this. A fitting mounted on the left hip and its mirror on the right experience different twist directions when the wearer rotates, and the half that is loaded in the opening direction is the one that will complain first. Where a design uses both sides, test both sides rather than testing one and assuming symmetry.

Mounting direction relative to the load is the last piece. Put the release direction across the body rather than away from it where possible, so that normal movement presses the halves together instead of prising them apart. This is a layout decision made on the pattern, and it costs nothing to get right and a season of returns to get wrong.

Where a fitting must be mounted on a curved or wrapping face, build the sample on the curve rather than flat. A housing seated while the panel is flat is loaded differently once the panel is wrapped round a torso, and the preload can either help registration or pull the halves out of it. Making the decision on a flat bench and discovering the problem on a body is the most avoidable failure in this category.

Bottom line: Fix handedness and orientation cues in tooling rather than in instructions, since a panel that assembles one way only removes an entire class of assembly errors from every shift that follows.

Reading the datasheet: five lines that matter

Datasheets in this category vary in quality, and the missing lines are usually the ones that decide field behaviour. Ask for five items as a minimum and treat any gap as a question rather than as a minor omission.

Note the third row specifically. A temperature range is not a curiosity: polymers stiffen in the cold and magnet grades lose field in the heat, and a fitting specified without a range cannot be warranted for a product sold across seasons and regions.

Ask also for the mating panel preparation the figures were measured on. A force measured on a rigid test plate will not be reproduced on a soft panel, and the difference is often larger than the margin between two competing candidates.

Where two candidates are close on paper, decide on the datasheet's completeness rather than on the highest number. The part with a stated temperature range, a stated stack envelope and a stated cycle figure can be specified and enforced; the part with one impressive number and four omissions cannot, and the omissions always surface after the order.

Datasheets omit what marketing does not need, so ask separately about coating durability on plated parts and reference behaviour under ASTM B117 rather than trusting a corrosion-resistant adjective.

Integrating the fitting into a sewn panel

Integration is mostly about giving the fitting something rigid to sit on. A local stiffener behind the housing keeps the two halves in register through wear and washing; without it, the panel flexes, the halves separate slightly and self-location degrades. One extra layer and one operation is usually enough, and the layer should be captured by the same stitch block that carries the housing.

Sequence matters. Fit the housing after the panel is stabilised and before any operation that could distort it, and use a locating jig so every unit places it identically. Operator instruction should state the sequence explicitly, because the common fault is not a broken part but a rotated housing on a subset of units.

Then verify the stack on the finished unit rather than on the components. Measure total thickness under the flange on production samples, measure the offset between the halves, and record both in the first-article report. Where the numbers drift outside the stated envelope, the fix is in cutting tolerance or stiffener specification, not in asking the assembly team to press harder.

Cleaning and laundering deserve a line of their own. If the product is washable, the panel will shrink and the stiffener will relax differently from the shell, and the register the fitting depends on will move. Run ISO 6330 on assembled units, re-measure the offset afterwards, and set the acceptance envelope from the post-wash figure rather than the pre-wash one. Where the product is not washable, say so on the care label instead of leaving the customer to find out.

Washable claims need care-cycle evidence, so confirm behaviour against ISO 6330 and write the allowed method onto the care label instead of letting customers experiment at home.

Failure signatures unique to this principle

Every closure fails, but magnetic-guided snap fittings fail in ways that are easy to misdiagnose. Recognising the signature saves a round of sampling.

The third row is the expensive one, because it appears after the product has shipped and looks like a quality problem rather than a materials decision. A stiffener chosen for thickness rather than for creep resistance will hold its dimension on day one and lose it by month six. Specify the stiffener by material and by compression behaviour, not by thickness alone.

Selection procedure for a programme

Start from the load case, not from the part. Write down the mass the module will carry, the direction the load acts, how often the fitting is cycled, and the temperature range the product will see. Those four items define the acceptance criteria; everything after is matching.

Then shortlist by motion rather than by brand. Decide whether a straight push, a slide, a tip or a rotate suits the user and the snag risk, and filter candidates by that. Only then look at force figures, and compare them on a mating panel preparation that resembles yours.

Finish with the commercial and process block: part number and frozen footprint, 25 mm webbing mounting with 38 mm vertical spacing and 50 mm horizontal repeat where the chassis uses a webbing field, permitted stack and offset, reinforcement and jig requirements, incoming inspection at AQL 2.5 under ISO 2859-1, MOQ 500 per variant, sampling 6-10 working days, screens and tooling at USD 300-2,500, and mass production 35-50 days. A selection that ends with those lines written down is a specification; one that ends with a preference is not.

One last decision belongs in the same document: whether the footprint is exclusive. If the drawing is built around one supplier's geometry, the programme has accepted a single source on price, colour and lead time. That is a legitimate choice, but it should be recorded with a mitigation - either a second qualified part engineered to the same footprint, or an agreement on notice periods for discontinuation. Hardware lines are discontinued with less warning than most buyers expect.

Judgement: Approve integration only after cycle testing, panel pull testing and one disassembly trial succeed together, then record those acceptance numbers on the order rather than in the development correspondence.