Tensile testing for small parts: fixtures, specimens and results

What a tensile test measures

A tensile test pulls a specimen until it yields and then breaks, recording load and extension. From the curve comes the yield strength at a defined offset, usually 0.2 per cent for metals, the tensile strength or maximum load divided by the original area, and the elongation after fracture as a percentage of the gauge length. For some materials the test also gives the reduction in area and the modulus in the elastic region.

Those numbers describe the material in the direction of the test, at the rate of the test, on a specimen of that size and surface condition. Change any of those and the number can move. That is not a reason to distrust the test. It is a reason to record the conditions and to compare like with like.

Why small parts behave differently

A small specimen has a larger surface to volume ratio, so surface condition and edge quality matter more. A sheared edge from a stamping, a grinding mark or a plating layer can start a crack earlier than it would in a large bar. The measured elongation falls, and the strength may rise slightly because the section is thinner and the strain path is different.

The size effect is real but not simple. A thin strip often shows higher strength and lower elongation than a thick bar of the same heat. Grain size relative to the section thickness also matters: with only a few grains across the thickness, a single soft grain can dominate the result. Report the specimen dimensions so the reader can judge whether the number is comparable with the material certificate.

Specimen options for small parts

Where the part geometry allows, test the full section. A wire, a small pin or a narrow strip can be gripped at both ends and pulled without machining a reduced section. This tests the real surface and the real edges, which is often what the designer wants to know.

Where the full section cannot be gripped, a sub-size or miniature specimen with a reduced parallel length and generous fillets can be machined or punched. The risk is that the specimen no longer represents the part, particularly if the working section is taken from a region with a different grain flow. A coupon from the same strip is a compromise: it represents the material lot, not the formed part.

Gripping without introducing the failure

The grip is the most common source of invalid results on small parts. Wedge grips concentrate load at the jaw face and can crush a thin strip or nick a wire. A capstan or bollard grip wraps the wire around a drum and loads it by friction, which avoids jaw marks but needs enough wrap length to prevent slip.

For very small or brittle parts, potting the ends in a resin block gives a uniform grip and spreads the load. The trade-off is preparation time and a gauge length that includes some of the potted region. Whichever method is used, check that the failure occurs in the parallel length. A break at the grip is a fixture result and should be reported as invalid, not averaged into the batch.

Alignment and extension measurement

Bending is the quiet killer. A specimen that is not aligned with the load axis sees a stress gradient across the section, yields earlier on one side and breaks at a lower load. A spherical seat or a self-aligning grip helps, and a careful check with a strain gauged specimen confirms the alignment before a production series.

Extension can be measured from the crosshead or with an extensometer. Crosshead travel includes machine compliance, grip seating and any slack, so it overstates extension, especially on a stiff small specimen. An extensometer clipped to the gauge length gives a much better modulus and yield, but its weight and knife edges can affect a thin specimen. Use the lightest extensometer that suits the section, or measure the gauge marks after the break for elongation only.

Rate control and its effect

Metals are rate sensitive. A faster test generally raises the measured strength and can lower the elongation, particularly in materials that show a yield point or a strain ageing behaviour. The standard therefore defines a rate, either as a strain rate through the elastic and plastic regions or as a stress rate, and the report should say which was used.

On a small specimen the machine control matters. A stiff frame and a closed loop control on extensometer strain holds the rate better than a constant crosshead speed, which produces a strain rate that changes as the specimen yields. Where the contract calls for a specific method, confirm that the machine and the extensometer can hold it before the test, not after the curve has been produced.

Reading the result

Report the individual results, the number of specimens and the mean, with the scatter. Three specimens that give 520, 545 and 610 MPa describe a different material from three that give 550, 552 and 555 MPa, even if the means are close. The scatter is often the most useful information about a small-part process.

Note where each specimen broke. A result that comes from a break near a fillet or a grip is suspect, and the standard usually allows a retest only under defined conditions. Check the fracture face for a defect before accepting a low value as a material property, because an inclusion, a fold or a plating crack will produce a low result that says nothing about the bulk metal.

• Specimen type, dimensions and gauge length.
• Test standard, rate and control mode.
• Individual results, mean and scatter.
• Break location and fracture appearance.
• Grip type and any slipping or jaw marks observed.
• Direction relative to the grain or rolling direction.