What does a clean sample of 14 prove?

That the defect rate is probably below about 19% at 95% confidence. It detects gross failures such as a lamp out, and it supports no PPM claim.

How many samples would a PPM target need?

About 15,000 inspected with no defect for 200 PPM at 95% confidence, more than six shifts of full inspection. That is why sampling cannot prove PPM quality.

Does higher confidence mean a better result?

No, it means a stronger claim from the same evidence, so the bound is weaker. At 99% the same 14 samples only rule out 28%.

If sampling cannot reach PPM, what should?

Monitoring the cause: dose on the part, radiometer trends, lamp hours and belt speed. Those cover every part and catch a drifting lamp early.

UV Curing calculator

UV Cure Validation Sample Size and What It Proves

Size a cure validation pull, then see what a clean sample proves. Enter shift volume, sampling rate, inspection efficiency and confidence level.

What this calculator does

Size a cure validation sample and state the defect rate it can rule out when every part passes.

Formula used

Raw samples = production volume × sampling rate ÷ 1,000
Sample pull = CEILING(raw samples ÷ inspection efficiency)
Defect rate ruled out = 1 − (1 − confidence)^(1 ÷ pull)
Samples for 1% = CEILING(log(1 − confidence) ÷ log(0.99))
Share inspected = sample pull ÷ production volume × 100

Inputs explained

Production Volume per Shift: Parts cured in the shift the validation covers.
Sampling Rate: Samples pulled per thousand parts.
Inspection Efficiency: Share of pulled samples that yield a usable result.
Confidence Level: Confidence for the upper bound; 95% is usual.

How to use the result

Best suited to sizing a routine validation pull, stating what a clean validation proves, justifying dose monitoring as the cure control.
One defect found moves the bound a long way; this page does not compute that. Assumes perfect inspection; undercure is exactly what a quick check misses. Says nothing about when defects occur within a shift.

Common questions

What does a clean sample of 14 prove? That the defect rate is probably below about 19% at 95% confidence. It detects gross failures such as a lamp out, and it supports no PPM claim.
How many samples would a PPM target need? About 15,000 inspected with no defect for 200 PPM at 95% confidence, more than six shifts of full inspection. That is why sampling cannot prove PPM quality.
Does higher confidence mean a better result? No, it means a stronger claim from the same evidence, so the bound is weaker. At 99% the same 14 samples only rule out 28%.
If sampling cannot reach PPM, what should? Monitoring the cause: dose on the part, radiometer trends, lamp hours and belt speed. Those cover every part and catch a drifting lamp early.

Last reviewed 2026-10-01.