Contact inspection: silver content, hardness and shear

What a certificate does not prove

A material certificate arrives with a lot number and a chemistry table. It does not tell you whether the face on the finished part is the material that was ordered, whether the sintering route produced the density the drawing assumed, or whether the joint to the shank will survive a pull test. Those are process questions, and they need process checks.

Contact inspection usually comes down to three core measurements: silver content, hardness and joint strength. A cross-section adds the internal view that the surface checks cannot show. Each check answers a different question, and each has a blind spot. A buyer who treats one as a substitute for the others is trusting a number that was never meant to carry the whole duty.

Silver content: what the assay actually measures

Silver content is a composition check. It confirms that the head alloy matches the grade on the drawing, within the tolerance the standard allows. The method matters. X-ray fluorescence is fast and non-destructive, but it reads the surface, so a plated or oxidised face can give a misleading result. Wet chemical analysis or inductively coupled plasma gives a bulk figure, at the cost of destroying the sample.

For a composite contact, sample the head, not the whole rivet. The shank has its own composition and will dilute the result if the two are digested together. Record the sample mass, the method and the uncertainty. A result of 89.6 per cent silver is not the same statement as 90 per cent plus or minus 0.5, and a drawing that calls for 90 per cent needs the tolerance defined before the first lot is argued about.

The oxide system matters as much as the silver figure. Two grades with the same silver content can behave differently because one uses tin oxide and the other uses nickel. The certificate should name the full designation, not just the silver number.

Hardness: a process fingerprint

Hardness on a contact face is a process fingerprint. It reflects density, sintering temperature, working and the distribution of the second phase. A reading that is too low suggests porosity or incomplete sintering. A reading that is too high can mean a brittle face that will crack under thermal cycling.

Use microhardness on a polished cross-section rather than a bench tester on a curved face. The indentation needs a flat, prepared surface and a known load, and the location should be stated: face centre, face edge or bond line. Vickers hardness at a small load is common for contact materials. Convert to another scale only when the conversion is valid for that material, because hardness conversions are not interchangeable across alloy families.

One indentation is an anecdote. Take a traverse across the face and a few points through the thickness, then compare the pattern with the first article. The absolute number matters less than the consistency between lots.

Shear and pull: testing the joint

Shear and pull tests answer different questions. A shear test loads the joint across its plane and is the usual check on a brazed or welded head-to-shank bond. A pull test loads the assembly along the axis and is the usual check on a staked rivet in its carrier. Using one where the other belongs produces a number that cannot be compared with the design intent.

The fixture decides the result. The sample must sit square, the load must pass through the intended axis, and the test speed must be stated. For a shear test on a small contact, a self-aligning fixture and a slow rate reduce the scatter that comes from off-axis loading. Report the failure location: bond line, head, shank or carrier. A strength figure without a failure mode is half a result.

Sample from each lot and keep the tested pieces. A photograph of the fracture face is often more useful than the number, because it shows whether the bond was continuous or only touching at the edges.

The checks and their blind spots

The checks below cover composition, structure and joint strength. None of them proves that the part will switch a load for a stated number of operations. That is a life test, and it belongs in the qualification plan rather than in a lot release.

Sampling that stands up

Sampling should follow the lot structure, not the calendar. A lot is a group of parts made from one material batch, on one line, in one run. If the material batch changes, start a new lot and a new record. For a new part or a new supplier, test more parts than the routine frequency, and include at least one part from each cavity or head position where that information is available.

A practical routine for a mature contact might be a small number of parts per lot for dimensions and hardness, with shear and cross-section on a lower frequency and on every change of material or process. Those frequencies are illustrative. The drawing, the application and the supplier control plan should set the real numbers.

Keep the untested remainder of the sample with the report. If a field failure appears later, the retained parts are the only way to compare the shipped condition with the current process.

Reading a result and reacting

An out-of-spec result needs a decision, not just a rejection note. First confirm the measurement: check the calibration, the fixture and the sample preparation. A hardness result that is 20 points low across a whole lot is a process signal. One low point on one part is more likely a pore or a preparation artefact.

Then decide the disposition. A silver content below the drawing limit is a material nonconformance and cannot be reworked. A hardness result outside the band may be accepted only with an engineering concession that states the effect on life. A shear result below the limit is a joint failure, and the lot should be contained until the cause is known.

Finally, close the loop with the supplier. Ask what changed: a new material batch, a sintering furnace, a plating line, a staking tool. A corrective action that does not name the change is a promise, not a control.