Glossary term
As-Found and As-Left
Engineering definition of as-found and as-left calibration condition covering adjustment records, out-of-tolerance impact, acceptance margin and release evidence.
Definition
methodAs-found and as-left describe the measured condition of an instrument before and after calibration, adjustment or service.
The as-found condition shows whether the instrument was still within acceptance limits when received for calibration. The as-left condition shows whether it was returned to service within limits after confirmation or adjustment. Together they support impact assessment, interval review, traceability records and release decisions.
As-found and as-left describe the measured condition of an instrument before and after calibration, adjustment or service. The pair matters because calibration is not only about returning an instrument to service. It also asks whether previous measurements made with that instrument were trustworthy.
The as-found result is the condition on arrival. The as-left result is the condition after adjustment, repair or confirmation. A good record preserves both rather than only stating that the instrument passed at the end.
As-Found Error
The as-found error compares the incoming reading with the reference value:
where (x_{AF}) is the incoming measured value and (x_{ref}) is the reference value. If (e_{AF}) is outside the acceptance basis, the instrument may have produced questionable historical data.
As-Left Error
After adjustment or confirmation, the as-left error is:
where (x_{AL}) is the post-service value. The as-left result supports release after calibration, but it does not erase the as-found condition. Historical impact still depends on how the instrument behaved before adjustment.
Adjustment Amount
The change caused by adjustment can be estimated as:
A large adjustment may be acceptable if the instrument is now stable, but it should trigger review of drift mechanism, calibration interval, affected measurements and whether a check standard showed the shift earlier.
Worked Example
A pressure channel is checked at a reference value of (100.00\ \text{kPa}). Its as-found reading is:
The as-found error is:
If the allowed error is (0.15\ \text{kPa}), the as-found result is outside tolerance before considering uncertainty. After adjustment, the channel reads:
so:
The instrument can be released as-left if the rest of the calibration record passes, but historical measurements need impact review.
Acceptance Margin
A conservative margin that includes expanded uncertainty (U) is:
where (T) is the allowed error magnitude. For (T=0.15\ \text{kPa}), (|e_{AF}|=0.18\ \text{kPa}) and (U=0.03\ \text{kPa}):
The negative margin means the as-found condition is not defensible for the stated tolerance without containment, correction, retest or risk acceptance.
Historical Impact
When as-found data fail, engineers should identify the affected interval:
where (t_{cal}) is the calibration date and (t_{last_good}) is the last passing check, maintenance event or credible measurement-control point.
The impact assessment should ask which results used the instrument, how close those decisions were to limits, whether check-standard data show a drift start, whether correction is possible and whether products, tests, reports or operations need containment.
If the error direction is consistent and the raw readings are available, a correction may be possible. If the instrument response is nonlinear, if the sign changed, if configuration records are missing or if decisions were close to limits, containment or retest is usually stronger evidence than a retrospective correction.
Impact review should also separate engineering use from administrative use. A monitoring trend used only for awareness may need annotation, while a release test, compliance report, safety interlock proof test or product acceptance record may need formal disposition.
Relation to Calibration Interval
As-found history is one of the strongest inputs to calibration interval review. Repeated clean as-found results can support interval extension when risk is low and use is stable. Repeated near-limit or failed as-found results support shorter intervals, intermediate checks, better handling or a more stable instrument class.
As-left history also matters. If the instrument repeatedly needs large adjustments to pass as-left, the problem may be drift, environment, handling, overload, firmware scaling or an unsuitable measurement principle.
Limits and Common Mistakes
Common mistakes include recording only the final pass, adjusting before measuring as-found condition, treating as-left acceptance as proof that historical data were valid, ignoring uncertainty in the impact decision and failing to notify process owners when an out-of-tolerance result could affect prior release decisions.
Another mistake is assuming all past measurements are invalid after one failed as-found result. The correct response is bounded impact assessment. A strong as-found/as-left record states the reference, readings, errors, uncertainty, acceptance basis, adjustment, affected interval, containment decision and release status.