Most of what goes wrong in e-bike wiring harness sourcing for two-wheeler brands does not show up at incoming inspection. It shows up eight months later, in a service queue, as a bike that intermittently drops the throttle signal in the rain, or a scooter whose battery lead has gone soft and discoloured behind the ring terminal. The harness passed continuity. It passed a pull test. It passed a visual. And it still generated a claim.
That gap between a harness that tests good and a harness that survives a product life is where two-wheeler brands lose money. A bicycle or a scooter is a hostile place for a wire: it is wet, it vibrates, it is made of sharp-edged tube and weld seam, its cable runs are long relative to the voltage, and its user hoses it down and then leaves it outside. Three failure modes account for the bulk of what a warranty department sees, and all three are decided on the drawing, not on the assembly bench.
Why two-wheeler harnesses fail differently
An appliance harness sits in a cabinet at controlled temperature and never moves. A vehicle harness runs in a body shell with tooled routing channels and a defined moulded path. A two-wheeler harness has neither: it is routed by hand along an extruded frame, through a head tube that rotates, past a fork that moves, into a battery cradle the user removes and refits several hundred times a year.
So it is loaded in ways your bench test does not reproduce. Movement, abrasion, water and heat arrive together and compound — a chafed jacket admits water, water sits in the connector cavity, corrosion raises contact resistance, resistance raises heat, heat degrades the seal further. A single defect rarely presents as a single symptom.
| Failure mode | What the rider reports | Where it was decided | What it costs you |
|---|---|---|---|
| Water ingress at a connector or splice | Intermittent throttle, display drop-out, phantom errors that clear when dry | Seal selection and the wire outer diameter it was sized for | Diagnosis time, no-fault-found returns, repeat visits |
| Chafe where the harness crosses the frame | Dead light, dead sensor, later a short to frame | Routing, clamp spacing, abrasion protection | Full loom replacement, often out of a service window |
| Battery lead sized on peak current | Warm or discoloured terminal, voltage sag on climbs, melted insulation | Conductor cross-section and insulation temperature class | Safety escalation, potential field action |
Read that table as a sourcing brief, not a failure catalogue. Every row is something you can put on a drawing.
The three failure modes that generate claims
Water gets in at the wire, not at the connector face
Buyers ask for a sealed connector and assume the problem is solved. The connector face is rarely the leak. Water arrives at the back of the housing, at the point where each wire passes through its individual seal, and it arrives from inside the wire as often as from outside it.
Two details decide this, and neither is usually written down.
First, the single-wire seal in a sealed housing is sized for a specific insulation outer diameter, with a working band typically a fraction of a millimetre wide. If your supplier substitutes a wire with the same conductor cross-section but a thinner wall — a change that looks free on a cost sheet and does nothing to electrical performance — the seal no longer grips. The assembly still passes continuity, insulation resistance and a withstand-voltage check. It fails in the first hard rain. Any wire substitution on a sealed circuit has to be re-qualified against the seal, not just against the current rating.
Second, stranded copper wicks. Cut a stranded conductor, expose it at a ring terminal or a spliced joint near the battery cradle, and water will travel along the interstices between strands, under the insulation, for a surprising distance — and it will emerge inside the connector at the far end, behind every seal you paid for. This is why sealed two-wheeler looms use adhesive-lined heat-shrink over splices and over the barrel of exposed ring terminals, and why an unsealed cut end anywhere on a wet circuit undoes the sealing on the rest of it.
When you specify an ingress rating, specify it properly. Ratings under the general ingress standard, IEC 60529, apply to a connector mated with its intended counterpart, correctly assembled with the specified wire and with every unused cavity plugged. An unmated rating is a different, and usually much lower, number. A supplier quoting a high water rating for a connector sitting loose in a bag is quoting something that does not describe your product. Note too that the higher-pressure washdown rating used in road-vehicle work comes from a different standard, ISO 20653, and is not interchangeable with the IEC number.
The frame is an abrasive tool
Aluminium frame tube, weld seam, cable exit hole, battery cradle lip, kickstand mount: each of these is a file that runs against your jacket every time the bike goes over a kerb. Vibration amplitude is small. Cycle count over three years is enormous.
The mistakes are consistent. Cable ties pulled tight over a bundle at a bend concentrate the load exactly where the jacket is already stressed, so the tie becomes the wear point rather than the restraint. Clamps placed only on one side of a frame pass-through let the harness saw against the hole edge. Service loops drawn generously in CAD get cut short on the line. And unprotected wire runs through a hole with no grommet, because the grommet was a line item that never made the bill of materials.
What works is dull and specific, and the choice between tape wrap, spiral wrap and corrugated conduit is worth making deliberately rather than leaving to the shop: an abrasion sleeve or conduit over every span that touches or could touch structure, a clamp on both sides of every pass-through, a designed service loop at the battery and the steering so the movement happens in the loop rather than at the connector body, and a stated maximum unsupported span — many two-wheeler drawings settle somewhere in the 150–200 mm range, though the right number depends on your frame and bundle mass. Put the number on the drawing rather than leaving it to the fitter.
One more detail that separates a considered harness from a copied one: the harness should be built so its natural resting shape matches the routed path. A loom taped straight and then forced around the head tube carries a permanent bending moment against the first clamp, and that clamp is where the jacket opens up.
A battery lead sized for peak current is sized wrong
This is the failure that turns into a safety conversation, so it is worth being blunt about the arithmetic.
Controller data sheets quote a peak: some number of amps for ten or twenty seconds. Buyers size the battery lead on it, feel conservative, and move on. But the thermal case is not a twenty-second peak in still air on a bench. It is a sustained draw on a long climb, inside a closed frame tube or a battery compartment that has already warmed up, in a bundle with other current-carrying wires, in an ambient that can sit well above room temperature. Continuous current in a confined, bundled, hot installation is the sizing case. Peak is a secondary check.
Three consequences follow.
The insulation temperature class matters as much as the copper. A standard PVC-insulated wire rated to 105 °C and a cross-linked or silicone-insulated wire of the same cross-section have very different headroom once the enclosure is already at 60 °C. Specify the insulation material and its temperature class, not just the cross-section.
The crimp, not the wire, is usually the hot spot. A joint with half a milliohm of contact resistance carrying 40 A dissipates about 0.8 W into a volume of a few cubic millimetres. That is why terminal cross-section analysis and pull testing exist as separate checks — the section shows whether the strands were compacted correctly and the pull test shows whether the mechanical grip is there, and a crimp can fail one while passing the other. Ask for both, on your wire and your terminal, not on a sample the supplier chose.
Voltage drop scales against system voltage, so a low-voltage bike suffers most. A one-volt drop on the run to the controller is under two per cent of a 52 V pack and nearly three per cent of a 36 V pack — and it lands as reduced assist at exactly the moment the rider is working hardest, which is what generates the complaint. Specify a maximum permitted drop across the battery lead at continuous current and let the supplier work back to a cross-section.
The termination style on the pack side follows from the same reasoning — the trade-offs between a bolted ring terminal and a quick-disconnect are as much thermal as they are ergonomic, and the conductor sizing and lug crimping detail behind a high-current lead deserves its own review.
Finally, coordinate the lead with its protection. Whatever fuse or battery-management trip protects that circuit has to clear before the lead's own thermal limit is exceeded. If nobody has checked that, nobody has checked the most important number on the harness.
What the drawing has to say
Suppliers build what you specify. If your specification is a photograph of last year's loom and a note saying "same as sample", every substitution decision moves to the supplier by default. Use this as the minimum block on a two-wheeler harness drawing:
- Circuit list — from-to for every conductor, with function, colour and printed wire marking.
- Wire specification per circuit — conductor cross-section in mm², strand class, insulation material, temperature class, and the insulation outer diameter tolerance where the wire passes through a seal.
- Continuous current per circuit, plus ambient and bundling assumptions, and a maximum permitted voltage drop across the battery lead.
- Sealing requirement per connector — required ingress level, stated as mated, the standard it is measured against, cavity plugs for unused positions, and adhesive-lined shrink over every splice and exposed terminal barrel.
- Routing and protection — abrasion sleeve or conduit by span, clamp positions and maximum unsupported span, grommets at every pass-through, service-loop length at battery and steering.
- Termination quality gates — crimp height with tolerance, minimum pull force by cross-section, cross-section sampling frequency, and who signs the first-article report.
- End-of-line test — continuity plus insulation resistance and a withstand-voltage check at a stated voltage, on 100% of assemblies, with the pass criteria written down.
- Change control — no substitution of wire, terminal, seal or sleeve without written approval, stated as a purchase-order condition. On sealed circuits, nominate the connector system by part identification rather than by description — that single line removes the most common silent substitution in the category.
For the wider version of this discipline, the CMH guide on writing a product spec sheet for a Chinese factory covers how to structure the document so it survives translation and internal handover at the supplier.
Where certification sits, and what the harness contributes
The harness is almost never certified on its own. In the EU, a pedal-assist electric cycle is assessed at the product level under the EPAC standard EN 15194, and the wiring, connectors and protection choices feed the equipment-level technical file rather than earning a separate cable approval. In North America, several retail channels, insurers and municipalities now require the battery and electrical system of an e-bike or scooter to be evaluated to a recognised system-level electrical safety standard; which one applies depends on your channel, and the harness is assessed as part of that system, not on its own.
Practically, that means two things. Ask your supplier for the construction evidence a test house will want — wire and insulation material declarations, terminal and seal specifications, and the test records behind them — rather than for a certificate on the harness. And when you are shown any certificate, check three things on the document itself: the issuing body, the holder's name, and the validity date. Certificates in this trade are issued per standard and per holder; one in a related company's name, or one that expired last year, is not evidence about the parts you are buying.
Common questions
Is 1,000 pieces a realistic first order for a custom two-wheeler loom?
For a build-to-drawing harness, a four-figure minimum is normal because the setup cost sits in applicator changeover, first-article approval and test-program creation rather than in material. Huisheng Electronics, for example, states a 1,000-piece minimum with samples dispatched in 7–10 days and volume output at 15–25 days to the warehouse — supplier-stated figures worth confirming against your own drawing before you plan a launch. Your room to negotiate sits in splitting that quantity across variants of the same base loom.
Can a small harness shop build a genuinely sealed loom?
The equipment question is narrower than it looks. Sealed assemblies need controlled crimping, correct seal seating, cavity plugging discipline and an end-of-line electrical test — not exotic machinery. What you should verify is whether the shop has built sealed assemblies before, what seal and cavity-plug hardware it stocks, and whether its outgoing test includes insulation and withstand-voltage checks rather than continuity alone. Ask to watch the seal-seating step on a live video call.
What does a salt-spray test actually tell me here?
Less than buyers assume, and something worth having. A salt-spray chamber shows whether plating and terminal finish resist corrosion under a standard exposure; it does not reproduce the wet-dry cycling with vibration and current that kills a two-wheeler connection. Treat it as one input among several, and pair it with a soak-and-power test on your own assembly if the circuit matters.
What to ask your harness supplier next
Send these with your drawing, and judge the supplier by which questions get a direct answer:
- What sealed connector systems do you stock, at what ingress level, and is that rating stated mated or unmated?
- What continuous current has this battery lead construction been validated at, at what ambient temperature, and in what bundle? If there is no validated figure, say so.
- What is your crimp height specification for this terminal and wire, what pull force do you accept, and how often do you cut a cross-section?
- Show me your end-of-line test: what voltage, what pass criteria, what percentage of units.
- Which abrasion protection and clamp hardware do you supply fitted, and which do I install?
- Which of the wire, terminal, seal and sleeve on this drawing would you propose to substitute, and why?
A shop like Huisheng Electronics in Longhua, Shenzhen is a reasonable shape for this kind of work — two-wheeler connector leads and battery looms with ring-terminal and quick-disconnect options are one of its six listed lines, and its published roster covers automatic cut-strip-crimp machines and servo crimping presses alongside a bench with cross-section, tensile, insertion-force, withstand-voltage, insulation, salt-spray and continuity testers. Capacity claims from any small Shenzhen shop are supplier-stated until you see them; a live walkthrough of the crimping and test benches, followed by a structured sample order and a pre-shipment inspection against your own written criteria, is what converts a plausible supplier into a qualified one.