Two wires swapped in a 40-way connector look perfect to the eye: same colour family, clean crimps, contacts fully latched. The harness passes a quick “beep” test because every wire conducts. It is only on the test bench that a sensor reads the wrong channel, and a day is lost looking for a software bug that is really a wiring error. Harness testing exists to catch exactly that before the harness leaves the workshop, and to catch the faults no one can see: a strand bridging two contacts, a damaged insulation, a crimp that will loosen under vibration.
This guide covers the electrical checks (continuity, shorts, insulation), the visual and mechanical checks that complete them, and how to organise them into a repeatable procedure.
1. The reference document: the wiring table
Every check starts from the wiring table (or netlist): for each wire, its start point (connector and cavity) and end point, and for splices, every point of the same net. Without a validated wiring table, a tester can only confirm that “something conducts”, not that the right things are connected.
Good practice: the wiring table is generated from the same source as the manufacturing drawing, versioned, and referenced in the test report.
2. Continuity, point to point
2.1 Principle
For each net, the tester injects a small current between the expected endpoints and measures the resistance. A pass means the resistance is below a threshold. A point-to-point test checks each pair of endpoints of the net, not just “one end beeps to something”.
2.2 Setting the threshold: worked example (illustrative values)
The threshold must be above the expected resistance of a good wire plus contacts, and below the value that would reveal a defect.
Example: a 2 m wire of 0.35 mm² copper. Copper resistivity is about 0.0175 Ω·mm²/m, so R = ρ·L/S = 0.0175 × 2 / 0.35 = 0.1 Ω. Add two crimped contacts of a few milliohms each: about 0.11 Ω. A threshold of, say, 1 Ω accepts all good wires of this length with margin, and still detects an open circuit. It will not detect a crimp that is half-made but still conducting at 0.3 Ω; for that, a lower threshold and a 4-wire (Kelvin) measurement are needed, or the mechanical pull test described in section 5. Set the threshold against the longest path in the harness, not an average wire.
2.3 Swapped wires
A swap is detected only if the tester checks that each endpoint connects to its expected partner and no other. This is why continuity and short tests are usually run together as a full netlist verification.
3. Shorts between lines
After continuity, every net is checked against every other net: no conduction is allowed between nets that the wiring table keeps separate.
- The threshold for a short is high (for example, any resistance below a threshold set in the test specification, often in the kΩ to MΩ range depending on the circuit, counts as a short), and is defined by the test specification.
- The number of combinations grows quickly: for N nets, N(N−1)/2 pairs. Example: 40 nets give 40 × 39 / 2 = 780 pairs. Testing them manually is unrealistic; this is the job of an automatic harness tester, which can test one net against all others connected together, then narrow down if a fault appears.
- Shields and drain wires are nets too: decide whether they must be isolated from everything or connected at one end.
4. Insulation under voltage
4.1 Why low voltage is not enough
A continuity tester works at low voltage. A damaged insulation or a contaminated connector may look perfect at a few volts and break down at the voltage the harness will see in service, or with humidity. An insulation test applies a higher DC voltage between each net and the others (or the shield, or the housing) and measures the leakage current, from which an insulation resistance is calculated.
4.2 Values come from the drawing, not from habit
The test voltage, the duration and the minimum insulation resistance are set by the customer drawing or the applicable standard. The guide cannot give a universal value, and nor should a test operator.
Example (illustrative values only): a test at 500 V DC with a minimum of 100 MΩ means the leakage current must stay below I = V / R = 500 / 100 × 10⁶ = 5 µA. The tester must therefore resolve microamps reliably, and the setup (fixtures, test leads, humidity) must not itself leak at that level.
4.3 Safety
Insulation testing involves voltages that can be dangerous. The fixture should prevent touching contacts under test, the voltage must be discharged before handling, and the operator procedure must say so explicitly. The test is performed only by trained, authorised staff.
5. Visual and mechanical checks
Electrical tests do not see everything. A crimp can conduct perfectly today and fail after vibration. The visual and mechanical checks are therefore part of the test, not an option.
- Stripping length and quality: no nicked or cut strands, correct length for the contact.
- Crimp: conductor and insulation crimp in the right zones, bellmouth present, no strands outside the barrel.
- Contact latching: each contact is fully seated and locked; a light pull-back from the rear confirms it. Secondary locks engaged.
- Labelling: wire and connector markings match the drawing.
- General appearance: no damaged insulation, correct routing and fixing, protective sleeving in place.
As a reference for acceptance criteria, many manufacturers use IPC/WHMA-A-620 (requirements and acceptance for cable and wire harness assemblies), published by IPC. We use it as a good-practice reference; the customer drawing remains the governing document.
5.1 Crimp pull test
A crimp pull test checks the mechanical strength of a crimp by pulling the wire out of the contact until it breaks or reaches a minimum force. The minimum value comes from the contact manufacturer for the given contact and wire section. The test is destructive, so it is done on sample crimps made with the same tool and settings at the start of a batch and after any tool change. Crimp quality depends on the right contact, wire and tool combination (Crimping and contact selection for circular and automotive connectors).
6. Organising the checks into a procedure
At RAYNOV, cable and jumper assemblies are checked against a documented inspection sheet covering, in order: checks before manufacturing, length, stripping and crimping, pull test, contact latching, continuity, short circuit, insulation, labelling and appearance. Limits on that sheet come from each customer drawing. Whatever the format, a useful procedure:
- references the drawing and wiring table versions;
- lists each check with its method, threshold and source of the threshold;
- records the result per harness (serial number or batch);
- keeps the pull-test samples traceable to the tool and date;
- defines what happens to a non-conforming harness (repair, retest, scrap).
7. Common mistakes
- Testing continuity “end to anything” instead of point to point against the wiring table.
- A continuity threshold so high it hides poor crimps.
- No insulation test because “it is a low-voltage harness”.
- Insulation values chosen by habit rather than from the drawing or standard.
- Skipping the pull test when the crimp tool or reel of contacts changes.
- Visual inspection done after the harness is closed in sleeving.
8. Checklist before delivery
- [ ] Wiring table version matches the drawing
- [ ] Point-to-point continuity, threshold documented
- [ ] Shorts between all nets, shields included
- [ ] Insulation test at the voltage, duration and limit from the drawing
- [ ] Stripping, crimping, latching and labelling inspected
- [ ] Pull-test samples within the manufacturer’s minimum
- [ ] Results recorded per harness, non-conformities closed
Harness testing on benches and EOL stations
The same checks scale from a single cable on a bench to automated testing in an end-of-line station, and a tested harness is also the starting point of any reliable test bench. For a bench harness, the pinout check is part of acceptance (Qualifying a test bench before production (FAT/SAT)). Breakout access to a harness during development is covered in Breakout boxes: what they are for and how to specify one. RAYNOV manufactures harnesses and builds test equipment within its custom validation systems; acceptance testing is discussed in our article ATP vs QTP.
Need harnesses built and tested, or a tester for your own production? Contact us.
