Pull test and weld shear test: fixtures, samples and records

Two tests that answer different questions

A pull test loads a joint along the axis it will see in service and records the load at which it yields, slips or breaks. A weld shear test loads the joint across the interface between two joined members, so the weld itself carries the load in shear. Both produce a number in newtons, and both are often filed under the same heading on a test report. They are not interchangeable.

Use a pull test when the question is whether a termination holds: a crimped lug on a cable, a wire welded to a terminal, a loop joined to a formed wire part. Use a weld shear test when the question is whether the weld metal and the bonded area are large enough: a wire welded to a flat tab, a contact welded to a carrier, a stud welded to a plate. A part can pass one and fail the other, because the load path through the joint is different.

Setting up a pull test

The value of a pull test depends on the load being applied along the intended axis, with no bending moment at the joint. A specimen clamped slightly off centre fails at a lower load, and the report blames the part rather than the fixture.

Hold the termination over its full barrel or full palm, not just at the tip. Hold the other end at the feature the drawing defines, which may be the conductor, a loop or a formed leg. Keep the free length stated in the test method, because a short free length stiffens the assembly and moves the failure away from the joint. Apply the load at a controlled rate, since a fast pull adds dynamic effects and a slow one allows creep in soft copper or aluminium.

• Load cell calibrated and in date, with the expected failure between 20 and 80 per cent of full scale.
• Self-aligning grip or spherical seat so the axis follows the specimen.
• Free length recorded with the result.
• Crosshead rate stated in millimetres per minute.
• Failure photographed before the specimen is disturbed.

Setting up a weld shear test

A lap shear test pulls two overlapped members in opposite directions, so the weld carries load across its face. The fixture has to hold both members in line and keep the load in the plane of the weld. A peel component introduced by a misaligned fixture lowers the result and changes the failure mode from weld pull-out to parent metal tearing.

For a wire welded to a tab, the fixture grips the tab and pulls the wire in the plane or at a right angle. A right angle pull measures the joint in peel, which is the weaker direction and often the more realistic one for a wire harness. State which direction the test uses, because the two numbers cannot be compared. For a spot or projection weld between two flat parts, a chisel or a dedicated shear fixture loads the nugget across its diameter.

How many pieces to test

Sample size should follow the consequence of a joint failure and the maturity of the process. A new tool, a new weld setting, a new operator or a new material lot deserves more pieces than a stable process that has run for a year. Five pieces per lot is a common habit, and a habit is not a rationale.

For a critical joint, test at the start and end of each shift or lot. Destructive tests consume parts, so the plan belongs in the quote and the control plan, not in a conversation after the first shipment. The table below is an illustrative starting point rather than a standard.

The fixture decides more than the load cell

Most disagreements about a pull result come down to the fixture. A grip that slips shows a low peak and a long plateau. A jaw that bites into the wire creates a stress concentration, and the specimen breaks at the jaw instead of the joint. A fixture with a fixed centre that does not match the specimen pulls the joint into bending.

Check the fixture with a known specimen before blaming the part. A simple check is to test a length of the same wire with no joint at all. It should break in the free length at a load consistent with the material and the diameter. If it breaks at the grip, the jaws are marking the wire and every joint result from that setup is suspect.

Recording a result that can be audited

A peak load alone is a weak record. The report should show what was tested, how it was held and how it failed. The failure mode is often more informative than the number: a weld that pulls out of the nugget at a low load points to a different correction than a weld that tears the parent metal at a high load.

Keep the load displacement curve where the equipment provides one. The shape shows whether the joint yielded, slipped or broke suddenly, and a plateau before the peak suggests grip slip. Attach a photograph of the failed specimen with the fracture face visible. Link the record to the lot, the material certificate and the weld or crimp setting so that a later investigation can reconstruct the condition.

• Part number, drawing revision and lot or serial number.
• Test method, with the clause that defines the rate and the free length.
• Fixture identification and load cell calibration date.
• Peak load, yield load if recorded, and the unit.
• Failure mode from a controlled list, not free text.
• Operator, date and machine recorded with the result.

Where these tests stop

A pull or shear test is a single event at room temperature. It says little about fatigue, thermal cycling, corrosion or vibration. A joint that passes can still crack after repeated thermal cycles between two materials. Add a cycling test when the service condition includes movement, heat or a corrosive atmosphere.

Treat the pull and shear results as release checks, not as life predictions. The life question needs its own test: a bend or vibration cycle for a flexible joint, a thermal cycle for a bimetallic joint, a salt spray for a plated joint. Pair the mechanical result with inspection of the weld or crimp geometry, because a good load can hide a poor bond.