How to Read a Multimeter Calibration Certificate: As-Found vs As-Left

A multimeter returns from the laboratory with a new sticker and a report marked “Pass.” On a later page, an as-found result is outside tolerance. Has the meter passed or failed? Both statements can be correct, because they describe different points in the calibration process.

The practical questions are whether the meter is suitable for its next job and whether earlier measurements need review. Reading the certificate in that order turns it into a working quality record, rather than a document filed only for an audit.

Read the identity and scope before the result

Match the manufacturer, model, serial number and asset number to the instrument in your hand. Similar-looking meters are easy to exchange in a tool room. Check the calibration date, report revision and the laboratory that performed the work.

Next, find the functions and ranges actually tested. DC voltage, AC voltage, current and resistance are different capabilities. A certificate covering selected DC voltage points does not establish performance on every AC frequency or resistance range.

Compare that scope with the tasks the meter performs. A team checking a low-voltage control signal needs different coverage from a team measuring AC supplies. An instrument with additional insulation-testing functions may need a separate method and capability. Record exclusions instead of letting a broad model name imply complete coverage.

The EK-F598 product page, for example, lists several measurement functions. A calibration order should identify the functions and ranges required for your work; buying a multifunction instrument does not determine the calibration scope automatically.

As-found and as-left answer different questions

As-found data describe the instrument's performance before adjustment. As-left data describe its performance after any adjustment or repair and subsequent verification. If no adjustment occurs, the report may state that the as-left condition is the same as found. NI explains this distinction in its verification and adjustment guidance; the exact report format depends on the laboratory and procedure.

An as-left pass can support return to service within the tested scope. It does not erase an as-found failure. The latter is evidence about the meter's earlier condition and may trigger a review of previous work.

Do not assume as-found and as-left use identical limits. Some manufacturer procedures use tighter adjustment targets. Read the actual specification and decision criteria rather than comparing colored pass/fail labels alone.

Follow one test point from input to decision

Start with a row near a value used in your work. Identify the applied reference, selected range, instrument indication, reported error, tolerance and uncertainty. For AC measurements, also identify frequency and waveform conditions where given.

The following numbers are an illustrative example, not an EK-F598 specification or an actual certificate. A reference source applies 10.000 V and a meter indicates 10.012 V. If error is defined as indication minus reference, the error is +0.012 V, or +12 mV. The corresponding correction, under that definition, is -12 mV.

Now suppose the agreed acceptance limit is ±10 mV and expanded uncertainty is 2 mV. A simple comparison of the reported error with the limit places +12 mV outside the limit. The laboratory's formal conformity statement must follow the agreed decision rule. Uncertainty should not be subtracted ad hoc to turn the row into a pass.

Near a boundary, the rule matters. Another hypothetical result of +9 mV with 2 mV uncertainty can be treated differently under simple acceptance and a stated guard-band approach. Ask the laboratory to explain its rule before using the label to release equipment.

Pay close attention to units. A result expressed in percent, millivolts or parts per million can describe the same error in different ways. Also distinguish a tolerance expressed as a percentage of reading from one involving range or display counts.

Uncertainty is not another name for meter error

The reported error estimates how the indication differs from the reference at that test point. Measurement uncertainty describes the uncertainty associated with the reported result. The coverage factor and stated coverage basis help explain how an expanded uncertainty was formed.

A small error with a large uncertainty may provide less decisive evidence near an acceptance limit than the same error with a smaller uncertainty. Conversely, several extra display digits do not establish a small uncertainty. Use the certificate's information and the intended measurement requirement.

Do not copy a single point's correction across unrelated ranges. Applying corrections in daily work requires a controlled method, suitable interpolation where justified and an uncertainty assessment. For ordinary field work, the practical outcome may be a clearly defined release scope instead.

For the measurement procedure itself, see Digital Multimeter Calibration: Procedure, Accuracy and Traceability. This article focuses on what to do with the resulting report.

A traceability statement needs a measurement chain

A phrase such as “traceable to NIST” is not a promise that every future reading will be accurate enough for every application. NIST describes metrological traceability as a property of a measurement result supported by a documented, unbroken calibration chain, with each link contributing uncertainty. Its traceability policy and FAQ also explain the limits of extending a calibration result to later measurements.

Check the stated references, method and uncertainty. If accredited calibration is required, verify that the laboratory's current scope covers the relevant quantity, range and capability. A logo on a cover sheet should lead to that scope; it does not replace reading it.

Keep traceability, accreditation and fitness for use as separate questions. A result can have a traceable chain yet be too uncertain for a tight acceptance decision. The working environment, accessories and measurement method can also differ from the laboratory conditions.

What to do with an as-found out-of-tolerance result

First preserve the complete report, including the data before adjustment. Establish whether the meter is now released for use, restricted or awaiting repair. Then assign the retrospective review to the person responsible for measurement quality.

Build a list of work that relied on the affected function and range. Start with asset-use records, job sheets, test reports and available intermediate checks. A resistance-range issue does not automatically invalidate every voltage reading. Equally, a broad “Pass” on the final page is not a reason to ignore the failed range.

The review needs to consider the error's direction and size, the acceptance margin in previous decisions, how the meter was used, and what evidence exists about when the condition developed. Unless supported by records, do not assume the drift began on the calibration date or progressed linearly since the previous calibration.

Where previous results were comfortably separated from their decision limits, the responsible reviewer may reach a different conclusion than for borderline results. If the necessary records do not exist, state that limitation. Rechecking accessible items or using other independent evidence may be appropriate under the organization's procedure.

Document the conclusion, affected work, supporting evidence, actions and approver. The purpose is to decide which past decisions remain defensible and what needs correction, not simply to produce a longer failure report.

A receiving checklist for the tool room

• Instrument identity matches the report and asset register.
• Required functions, ranges and AC frequencies are covered.
• As-found and as-left conditions are understood.
• Adjustments, repairs and restrictions are recorded.
• Tolerances, uncertainty and the decision rule are available where needed.
• Any out-of-tolerance impact review has an owner and a recorded status.
• The release label agrees with the actual permitted use.
• The next review or calibration date follows the organization's interval policy.

This is an editorial working checklist, not a replacement for an accreditation body's required form. Add fields needed by your laboratory, contract or quality system.

Order the evidence before sending the meter away

Request as-found data before adjustment, as-left data after adjustment, the required test points, uncertainties and the conformity rule. Agree on how the laboratory should handle a failure or a repair outside the original scope.

Include the instrument's important operating ranges and any intermittent symptoms. A fault that appears only after warm-up, with particular leads or at a specific AC frequency is easier to investigate when the laboratory knows about it.

Finally, use the results when reviewing the next interval. Stable history, workload, transport, environment and the consequence of an incorrect result all matter. The calibration interval and decision-rule guide covers that separate decision. A new sticker closes the service visit; the certificate explains what the organization can safely rely on next.