As-found vs as-left

As-found readings are taken before any adjustment, cleaning or repair; as-left readings are taken afterwards. A certificate that only reports as-left data tells you the instrument is good now, but nothing about the measurements you made with it since its last calibration.

Why as-found matters

• It is the only evidence of how the instrument performed during the previous interval.
• An out-of-tolerance as-found result triggers an impact assessment on affected measurements, products or decisions.
• As-found history is the basis for setting or adjusting the calibration interval.

Reading the two together

Small drift between calibrations with no adjustment needed supports the current interval. Repeated adjustment at each calibration suggests the interval is too long or the instrument is unsuitable. A large as-found error after a small as-left error at the previous calibration points to an event — damage, misuse, environmental exposure — rather than drift.

In practice

Ask the laboratory to report as-found data even when no adjustment is made (in which case as-found and as-left are the same). Keep both in the instrument record. ValiTrac's calibration data model records both explicitly so that interval and impact decisions can be traced to evidence.

Four histories, and what each one is telling you

Read as-found and as-left together across successive certificates and the instrument tells you what to do with it.

The impact assessment nobody budgets for

An out-of-tolerance as-found result is not a calibration finding; it is a nonconforming-work event with a look-back attached. The work is to establish when the drift plausibly began and then, for every measurement the instrument supported since, whether the error was large enough to change a reported value beyond its stated uncertainty or to flip a conformity statement.

Most of them will turn out to be unaffected, and saying so with the arithmetic shown is the deliverable. What makes the exercise bearable is interim-check data: a monthly ice-point chart locates the onset to within weeks instead of within the whole interval, and turns a twelve-month look-back into a one-month one. That is the real return on a few minutes a month.

Getting the data in the first place

• Ask for as-found data explicitly, and ask for it even when no adjustment is performed — in that case as-found and as-left are the same, and the certificate should say so rather than reporting one set with no label.
• Tell the laboratory not to adjust without contacting you. The decision belongs to whoever owns the measurements the instrument has been making.
• Watch for certificates that report only as-left values with no statement about adjustment. That is not necessarily dishonest, but it withholds the one thing you needed.
• Keep both sets in the instrument record, not just the latest certificate. The history is what sets the interval, and it only exists if somebody kept it.

A certificate reporting only as-left data tells you the instrument is good now and nothing at all about the period it was actually being used in. That period is the one your product decisions were made in.