Why two kinds of limits

A lab has two completely different kinds of limits that look alike on paper (both are just

numbers — a lower and an upper limit), but they **answer a different question about a different

thing**. Confusing them is the single most common source of mistakes — which is why this LIMS

keeps them on two separate screens.

An example that separates them

A lab scale:

Checking the scale: you place a known 10 kg reference weight. If it reads ~10 kg, the

scale is fine. → "Is my method working correctly?" That's a Control-Chart Limit (§7.7.1).

Weighing the customer's sample: you weigh the customer's actual sample and compare it to a

requirement (e.g. "must be under 0.05 mg/kg lead"). → *"Does the customer's result meet the

specification?"* That's an Acceptance Criterion (§7.8.6).

Same-shaped numbers, completely different meaning.

Acceptance Criterion (ISO/IEC 17025 §7.8.6)

It's the specification the CUSTOMER's result is judged against to issue a **statement of

conformity** (conforms / does not conform).

Answers: "Does the customer's result meet the requirement?"
You set one only when a real specification exists: a regulatory limit, a customer-agreed

limit, or a product standard.

It's evaluated under a decision rule (simple acceptance, guard bands, etc. — see

"Why decision rules and conformity statements").

Drives the CONFORMITY column in the measurements grid and the conformity statement on the report.

Important: most agronomic parameters (e.g. the Ca/Mg ratio, C/N, available boron) have no

conformity specification. They don't "pass/fail" — they are interpreted (low / adequate / high)

via the Ranges and the Observations/Recommendations. For those, the CONFORMITY column correctly

shows "—" (Not Applicable).

Control-Chart Limit (ISO/IEC 17025 §7.7.1)

These are the control limits (target ± limits, Shewhart-style) for your own control material —

the known sample you run regularly to prove the method is in statistical control.

Answers: "Is my MEASUREMENT PROCESS in control?"
It's about the lab's Control Sample, not the customer's sample.
Drives pass/fail on Control Sample results.
This is also where the Replicate RPD limit (§7.7.1f) lives — the maximum acceptable percent

difference between an original and its replicate.

At a glance

	Acceptance Criterion (§7.8.6)	Control-Chart Limit (§7.7.1)
Subject	the customer's result	the lab's control material
Question	"Does it meet the spec?"	"Is the process in control?"
When you set it	only if a real spec exists	for every method/element you monitor with QC
What it drives	CONFORMITY + conformity statement	Control Sample pass/fail + RPD check
Screen	Quality Control → Acceptance Criteria	Quality Control → Control-Chart Limits

What the CONFORMITY column shows

✓ Conforming — inside the Acceptance Criterion.
✗ Non-conforming — outside the Acceptance Criterion.
⚠ Indeterminate — inside the uncertainty guard band of the limit (guard-bands rule).
— Not Applicable — no Acceptance Criterion exists for the element (the normal case for

agronomic parameters). The limits actually applied appear beneath the icon, so the ✓/✗ explains itself.

The practical rule

If you're setting limits to check whether your method is working → Control-Chart Limit.
If you're setting limits to state whether the customer's result meets a requirement →

Acceptance Criterion.

Don't register your control material's limits as Acceptance Criteria: then every customer's result is

wrongly judged "non-conforming" against a limit that was only ever about your own control sample.