Expired Load Cell Calibration During Product Release
During a product release run, a technician reports peak tensile loads for five polymer coupons that are consistently 8% below the historical mean for the same formulation. The bench-top universal testing machine uses a 5 kN load cell whose calibration certificate expired three months ago. A quick check with a 2 kN deadweight gives a displayed value of 1.94 kN; at 0.5 kN the display reads 0.50 kN. The laboratory is at 29 °C, while the load cell was calibrated at 23 °C; the manufacturer lists a temperature effect of 0.02% of full scale per °C. The zero reading drifts by 0.004 kN over ten minutes. The load cell has a 5 kN range, and the coupons normally fail near 3.8 kN. Production wants the batch released today, and the quality manager asks whether the tensile results can be reported with a note explaining the expired calibration.
1. Assess the measurement evidence and decide whether the tensile results can be released with an explanatory note. Show the quantitative reasoning that supports your decision.
Read the worked answer
The check data show a span or nonlinearity problem, not a simple zero offset. At 2 kN the error is 1.94 - 2.00 = -0.06 kN, or -3.0% of reading. At 0.5 kN the error is essentially zero. A zero offset would shift both points equally; instead the error grows with load. The temperature effect is small: 0.02% of full scale per °C × 5 kN × 6 °C = 0.006 kN, and the zero drift is 0.004 kN over ten minutes. Neither explains a 60 N error at 2 kN or the 8% drop in coupon results.
The larger issue is traceability. The calibration certificate expired three months ago, so the load cell has no current demonstrated link to a reference standard. A note cannot restore that traceability, and the 3% check error is far larger than the typical acceptance criterion for a testing machine load cell. The 8% drop is also larger than any identified environmental effect, so the cause may be the load cell, grip alignment, specimen preparation, or test speed. Releasing the results with a note would overstate confidence and could allow nonconforming material to ship. The correct decision is to hold the results, quarantine the machine from release testing, and investigate before re-testing.
2. Outline the corrective actions and retrospective impact assessment, including what records you would require before re-testing.
Read the worked answer
Immediate actions: remove the machine from release testing, label it out of service, and arrange an as-found calibration across the working range. Record the as-found data, including zero, span, linearity, hysteresis, and temperature. Check the test fixture, grips, alignment, and crosshead speed against the method, because an 8% shift can also come from mechanical setup rather than the load cell.
Retrospective assessment: list every batch tested since the calibration expiry date. Use the as-found bias and uncertainty to estimate whether any result could have crossed a specification limit. Apply a risk-based rule: retest batches whose results were close to the limit or whose disposition depended on the tensile data. If any result was already reported, issue a notification and correct the record.
Before re-testing, require: a valid calibration certificate traceable to a national standard; an as-left calibration within the manufacturer's tolerance; a successful intermediate check with a known deadweight or proving ring; documented verification of test setup; and a nonconformance record with root cause and corrective action. Re-test the five coupons and any affected batch on the verified system, and do not simply apply a correction factor unless the calibration supports it.