Crimp Height, Pull-Force and Cross-Section: How Harness Crimps Are Actually Judged

A machine builder in Ohio pulls an intermittent fault out of a returned drive cabinet. One lead in a control harness measures tens of milliohms where it should measure almost nothing, and the fault only shows up once the panel is warm. That crimp passed its pull test at the factory. It was pulled to a number, the number sat inside the range, the lot shipped.

Wire harness crimp quality inspection rests on three separate pieces of evidence — crimp height, pull force and a cross-section — and they are not interchangeable. Each answers a different question about the joint, and a supplier who can only show you one of the three is asking you to take the other two on trust. Below: what each measurement proves, what it structurally cannot see, and what to write into the purchase order so the records exist before your first container rather than after your first field return.

Why a crimp is a cold weld, not a squeeze

A correctly made crimp is a solid-phase weld formed at room temperature. The terminal's conductor barrel is compressed onto the strands until the copper deforms plastically, the surface oxide layers on both the strands and the barrel fracture and are pushed aside, and clean metal is forced into contact with clean metal under residual stress. The interface that results has no oxygen path into it. That is what "gas-tight" means, and it is the whole point of the operation.

Get that weld and the joint's resistance sits close to an equivalent length of the conductor itself, and it stays there through thermal cycling because there is nothing inside for moisture and oxygen to attack. Miss it and you get a friction fit that looks identical from the outside. On day one it conducts. Then it cycles, the interface micro-moves, oxide creeps in, resistance rises, the joint heats a little more at the same current, and the rise accelerates. The failure that shows up eighteen months later in a customer's cabinet was decided in a fraction of a second on a crimp press.

That is why crimp inspection is unlike almost every other check on a harness. Continuity, hipot and insulation resistance tell you the state of the assembly on the day it is tested. Crimp evidence has to tell you about a metallurgical condition that will not express itself for a year. You cannot test your way to that answer at the end of the line — you have to control the process that produced it, and keep the record.

Crimp height: the process variable that proves the weld

Crimp width is fixed by the die. The anvil and the crimper form a cavity of set width, so the only dimension that moves with press adjustment is the height of the finished conductor barrel. That makes crimp height the single controlled process variable, and it is the number a serious harness shop lives by.

It is measured with a blade or point micrometer, across the top of the conductor barrel at roughly the middle of the barrel length, deliberately away from the flared ends and away from the cut-off tab, read to a hundredth of a millimetre. The terminal manufacturer publishes a nominal crimp height and a tolerance for every combination of terminal and conductor size in its application specification; tolerance bands are typically in the hundredths of a millimetre, not the tenths. The applicator is set and shimmed to hit that nominal, and everything downstream is an argument about whether it stayed there.

It does not stay there on its own. Crimp height drifts as the applicator's tooling wears, as the press shut height shifts with machine temperature across a shift, and it steps abruptly when someone changes an applicator, loads a new reel of terminals, or clears a jam. A shop that measures crimp height once at the start of a 20,000-piece run and files the number has measured the setup, not the run.

Two details separate people who have done this from people who have read about it. First, the insulation crimp has its own height and its own target, measured separately — an over-crimped insulation grip cuts through the jacket and creates a stress riser, an under-crimped one lets the wire pivot and fatigues the conductor crimp behind it. Second, crimp height is meaningless without the terminal part number written next to it, because the same wire size in a different terminal has a different target.

Ask for the crimp height log in the format the shop actually keeps it: terminal part number, conductor size, target height, tolerance, measured values, time stamp, operator. If that log has to be created for you, it did not exist.

Pull-force: necessary, and not sufficient

A pull test — tensile test, crimp pull-out force — clamps the wire and the terminal and pulls until something gives, recording the peak. Published minimum values by conductor cross-section exist in the workmanship and automotive cable standards, so there is a defensible acceptance figure for any wire size you are buying. It is destructive, it is quick, and it is the test most Chinese harness shops will offer you first because it produces a number that looks like proof.

Here is the part that matters, and it is the reason crimps that pass pull tests fail in the field. Plot pull force against crimp height and you do not get a straight line — you get a curve with a peak. As crimp height comes down from loose toward nominal, retention climbs. Keep going below nominal and retention keeps climbing for a while, then rolls over. The peak of mechanical retention sits below the crimp height where the electrical joint and the strand integrity are at their best. A shop that tunes its press by pull test alone will drift toward over-crimp, because tighter reads as stronger on the gauge, right up to the point where strands are thinned, work-hardened and cracked and the barrel wings have split.

So a strong pull result is compatible with a bad crimp. It is also compatible with a strand set that was nicked during stripping — a pull test on a fresh lead barely notices two or three severed strands, but the flex life of that lead is gone. And the acceptance criterion itself has a hidden variable: whether the insulation grip is included in the pulled length. Included, it inflates the number. Define it in writing.

Pull force earns its place as a check that nothing gross is wrong and as a lot-level record. It does not earn the right to be the only evidence.

The cross-section: the only view of what happened to the strands

Cut a finished crimp transversely through the middle of the conductor barrel, mount it, polish it, etch it, and put it under magnification. This is the only inspection that shows the inside of the joint, and everything the other two tests are blind to lives here:

• Strand deformation and void closure. Individual round strands should be deformed into a compacted mass with the boundaries largely closed. Visible voids between strands mean the cold weld is incomplete, whatever the pull gauge said.
• Compaction across the section. Under-compaction leaves gaps; over-compaction thins the barrel base and work-hardens the copper until it cracks.
• Barrel form. The wings should roll fully inward and meet or interleave correctly at the seam, not cross over, not penetrate down into the conductor bundle, not leave a gap at the top.
• Base and wing thickness after crimping, plus any cracking in the plating or the base metal.
• Symmetry. An off-centre section says the terminal sat crooked in the die — a setup problem, not a random defect, which means it is in every piece from that setup.

Two adjacent features are worth learning to look for because they are the fastest way to read a supplier's process discipline. Bellmouth is the small flare at the ends of the conductor barrel; it is deliberate, and it prevents a sharp barrel edge from shearing strands under vibration. Missing bellmouth is a tooling or setup fault. Brush — the short length of strand ends protruding past the conductor barrel on the front side — confirms the conductor was inserted to full depth and is gripped along the whole barrel. No visible brush and you cannot tell whether the wire bottomed out short.

One practical demand: a cross-section photograph is only evidence if it carries traceability. Ask that each section image be labelled with the terminal part number, the conductor spec, the date, and the crimp height reading of that same crimp. An unlabelled micrograph is decoration, and every harness shop on earth has three nice ones to send you.

Common questions

If a supplier owns a cross-section analyser, is the crimp problem solved?

No — owning the instrument and running it on a schedule are different facts. A cross-section analyser proves capability; a written sampling rule proves practice. The rule you want is sampling tied to events rather than to calendar time: at first-article, after every applicator change, after a terminal reel change, and at a stated interval through long runs. Ask which of those triggers exists in the supplier's own work instruction, and ask to see the last month of results, not a curated set.

What is crimp-force monitoring and should I insist on it?

Crimp-force monitoring instruments the press to record the force-versus-displacement signature of every crimp and flag pieces that deviate from a learned reference — a missing strand, a doubled wire, an insulation strand in the barrel, a partially inserted conductor. It gives you 100% in-process screening rather than sampled inspection. It is common on higher-volume automatic lines and much less common on benchtop presses. Treat it as a specific question to ask about a specific machine, not as something to assume from a general equipment list.

Do I need all three checks on every order?

No. Match the evidence to the consequence. A low-current consumer lead with a sealed cost target may reasonably run on crimp height logging plus lot-level pull tests. A servo, battery or vehicle-side lead where a rising-resistance joint means heat, a fault code or a warranty return justifies paying for cross-sections at defined trigger points and keeping them on file with the lot record.

What to ask the supplier next

Send these with the drawing, before pricing, and treat the answers as part of the quote:

1. Crimp height targets and source. Which terminal application specification are you working to, and what nominal and tolerance will you use for each conductor size on my wire list?
2. The height log. Show a real one from a recent job. Check that it records re-measurement after applicator and reel changes.
3. A written pull-force acceptance table by conductor size, with the insulation-grip convention stated, and the standard it is derived from.
4. Cross-section sampling rule in writing — which events trigger a section, how it is recorded, how long the images are retained.
5. Crimp-force monitoring — which presses have it, and does the flagged-piece count get reported per lot?
6. Stripping control — strip length tolerance, how nicked strands are detected, and what happens to a lead with a suspect strip.
7. First-article package — ask for crimp height readings, pull results and labelled sections on the sample build, not just a working sample. Samples that arrive without process records tell you about the sample only, which is the classic route into a good-sample, bad-production order. The wider version of this problem is covered in our guide to product inspection in China before shipment.
8. A live bench walkthrough on video — the press, the applicator, the micrometer, the tensile tester, the section mount — rather than a photo set. What to look for in that call is laid out in the one-day factory inspection guide and in the broader notes on quality control in China manufacturing.

As a worked example of how far a public listing gets you and where the questions start, look at Huisheng Electronics, a Shenzhen harness shop in Longhua District incorporated in 2016 whose company name, credit code, registration status and registered address were checked against the Chinese corporate registry in August 2026. Its supplier-published roster lists automatic cut-strip-crimp machines, servo crimping presses and benchtop crimpers on the assembly side, and on the test side a terminal cross-section analyser, tensile and insertion-force testers, withstand-voltage and insulation testers, a salt-spray chamber and continuity testers — more than twenty machines in total. Stated trade terms are a 1,000-piece MOQ, samples dispatched in 7–10 days and volume in 15–25 days to the warehouse.

That roster answers the capability question and leaves the process questions open: no published pull-force acceptance table, no stated cross-section sampling frequency, no confirmation of which presses carry crimp-force monitoring, and no machine-by-machine breakdown behind the headline count. Those are exactly the six items above. Capacity figures on any small Shenzhen supplier are supplier-stated until an audit or a live walkthrough confirms them, and on certificates of any kind the only durable habit is to read the document yourself and check three things — the issuing body, the holder name, and the validity date.