Cable lug pull test: sample size and acceptance

What a pull test proves, and what it does not

A pull test applies a tensile load along the axis of the conductor while the lug is held so that the barrel and palm cannot move. The peak load is recorded, and so is the place where the joint or the cable failed. That pair of results is the evidence. A peak load on its own is a number without a cause.

The test proves the mechanical strength of the joint that was tested, at the moment it was tested. It does not prove that the barrel is free of voids, that the electrical resistance is stable, or that the joint will survive vibration and thermal cycling. Those questions need a cross-section, a resistance measurement or a longer test programme.

The pull test also cannot stand in for the crimp height record. If the load is low, the height tells you whether the die closed too high, the barrel was too large or the conductor was wrong. Without the height, the same low load could be blamed on the operator, the material or the tool, and the investigation starts from nothing.

Set the test up the same way every time

The fixture matters as much as the load cell. Hold the palm so the load passes through the barrel axis, and grip the conductor far enough from the barrel that the jaws do not damage the strands inside the crimp. A pull that is offset from the axis adds bending, and the result falls before the joint has reached its real strength.

Control the rate of load. A rapid pull can give a higher peak on a ductile joint, while a slow pull gives the strands time to rearrange. Record the rate in the test method so that results from different shifts and different machines can be compared.

Check the load cell and the gauge before the programme starts, and keep the calibration record with the test method. The pull test is often the last check before a lot is released, so a drifted load cell can pass a weak joint without anyone noticing.

• Pull along the conductor axis, with the palm supported and the barrel free of side load
• Grip the conductor beyond the barrel so the jaws do not crush the crimp
• Use one stated load rate and record it with every result
• Keep the load cell calibration current and traceable

Where the acceptance value comes from

The acceptance value should come from the product standard, the drawing or a written customer requirement. IEC 61238-1 sets out compression and mechanical connector tests for power cables, and UL 486A-486B covers wire connectors and soldering lugs. Each standard defines its own method and its own way of expressing the result, so quote the clause rather than a number copied from another job.

A value that was valid for a 95 mm² flexible copper cable is not automatically valid for a 95 mm² stranded aluminium cable, or for a different barrel wall. The conductor class, the material and the barrel geometry all change the result. If no value exists for the combination, agree one with the supplier and record the basis: the sample, the die, the height and the failure location.

A minimum load is a pass or fail line, not a description of quality. A joint that passes at exactly the minimum has no margin. It is better to state a target window, a minimum and a preferred failure location, so that a joint which passes by a small margin is still visible as a risk.

Sample size by stage

One sample size for the whole programme is usually either too small for qualification or too expensive for routine production. Split the plan by the question being asked. At setup, the question is whether the die and the height can make a sound joint. At first-off, it is whether the production press is set like the setup press. In process, it is whether the joint is drifting. At qualification, it is whether the die covers the full conductor range.

A practical plan tests more joints at the start of a programme and fewer during a stable run, provided the height and the visual checks continue. If the process is new, or the conductor has changed, go back to the qualification sample rather than relying on the routine frequency.

Read the failure location

A conductor that breaks outside the barrel is usually the best result. The joint held, and the wire reached its own tensile limit. A conductor that pulls cleanly out of the barrel points to under-crimp, a barrel that is too large for the conductor, or the wrong die. The strands look polished where they slid.

A barrel that splits points to over-crimp, a wall that is too thin, or a die with a sharp edge. A barrel that opens along one side can also come from a press that is out of alignment. Strands cut at the bellmouth point to a die that is flashing or a bellmouth that was closed flat, and the joint may still pass a pull test while it fails in vibration.

Record the failure mode in words that another person can use: conductor break outside barrel, pull-out with clean strands, barrel split, strand cut at entry, palm deformation. A photograph of each mode in the work instruction is more useful than a paragraph of description.

• Conductor break outside the barrel: the joint exceeded the wire strength
• Pull-out with clean strands: under-crimp, wrong die or oversized barrel
• Barrel split: over-crimp, thin wall or a damaged die
• Strands cut at the bellmouth: die flash or a closed bellmouth

Record enough to investigate later

The test record should identify the joint as fully as the drawing does. Record the lug part number and lot, the conductor size and class, the die number, the crimp height, the press, the operator, the date and the load cell ID. Add the peak load, the failure location and whether the sample passed.

Keep the failed samples. A labelled failed joint is evidence that can be sectioned later, and it is often the fastest way to settle a disagreement about the cause. If the load is trending down while the height stays in the window, the die or the barrel lot has changed, and the sample is the place to start.

A simple chart of load against crimp height, with the failure mode marked, shows more than a table of pass marks. It shows the margin. A lot that passes every test but sits close to the lower limit is a lot that deserves a second look before the next shipment.

Decide what happens after a failure

The control plan should say what happens when a sample fails, before the first failure occurs. A common sequence is to stop the press, quarantine the parts made since the last good sample, check the crimp height and the die, make a new sample and test it. The failed joint is kept, not recrimped.

Do not fix a failed crimp by crimping it a second time. The barrel is already work-hardened, and a second closure can crack the wall or damage the strands. Cut the joint off and start again with the correct die and height. If the failure is a pull-out, check the die number and the conductor range before changing the height.

After a die change, a press repair or a new barrel lot, re-qualify with the setup sample. If the failure was a trend rather than a single event, retest the quarantine stock at the routine frequency and record the result against the original lot. Release only when the cause is understood and the new samples pass with margin.

The test plan in one page

A usable plan fits on one page and leaves nothing to memory. It names the standard or drawing clause, the fixture, the load rate, the acceptance value, the sample frequency and the failure response. It also names who signs the release and where the records are kept.

• The product standard or drawing clause that sets the acceptance value
• The fixture, load rate and load cell calibration requirement
• The sample size at setup, first-off, in process and qualification
• The failure modes to record and the photographs that define them
• The quarantine and retest rule after a failure
• The record fields, including die, height, conductor and lot