To read a pressure gauge calibration certificate, match the instrument and pressure mode first. Then check the reported error, units, uncertainty, and acceptance rule at the pressures you use. A pass statement needs to be read with the specification it refers to.
This guide works through a small example for a gauge reading in bar. It helps you review returned results; it is not an operating procedure for applying pressure.
For the broader review process, read Reading Calibration Certificates with As-Found and As-Left Data.
Match the certificate to the job.
Before calculating anything, put the certificate beside the asset record and calibration request. Use the following questions as a practical review aid, rather than a list of mandatory certificate fields.
| Check | Question for the reviewer |
|---|---|
| Identity | Do the serial number and asset reference identify the gauge returned to you? |
| Pressure mode | Does the report describe gauge, absolute, or differential pressure as required by the job? |
| Units and range | Are the reported units clear, and do the calibrated points represent the intended operating region? |
| Configuration | Was the relevant gauge, transmitter, or complete measuring assembly included? |
| Condition | Are results before and after any adjustment clearly distinguished? |
| Specification | Which document and revision define the allowed error for this review? |
For pressure-specific method context, EURAMET cg-17 addresses mechanical and electromechanical manometers, including rising and falling pressure series and uncertainty contributions. Check the method actually named by your laboratory. EURAMET calibration guidelines, No. 17.
Read a worked result table.
Illustrative data only. These invented teaching values do not represent a laboratory certificate, customer instrument, or recommended test-point sequence. Assume a digital gauge with a 0 to 10 bar range. The example defines error as indication minus reference.
| Reference | Rising indication | Rising error | Falling indication | Falling error |
|---|---|---|---|---|
| 2.000 | 2.030 | +0.030 | 2.050 | +0.050 |
| 5.000 | 5.080 | +0.080 | 5.090 | +0.090 |
| 8.000 | 8.040 | +0.040 | 8.070 | +0.070 |
At the 5.000 bar rising point, subtract the reference from the indication: 5.080 - 5.000 = +0.080 bar. Under this definition, the gauge reads high. The corresponding additive correction at this point is -0.080 bar.
That arithmetic does not establish a correction for the whole range. A reviewer would still need the applicable correction method and evidence for other pressures and conditions. Do not copy one point's correction into every production reading.
Keep rising and falling readings separate.
At the same 5.000 bar reference, the two indications differ by 5.090 - 5.080 = 0.010 bar. This is the directional difference for the pair shown. A single pair does not establish the gauge's maximum hysteresis over its range.
EURAMET CG-17 includes hysteresis in the uncertainty treatment when rising and falling results are combined. Ask what a reported mean represents before using it for a process that approaches pressure from one direction. EURAMET CG-17, uncertainty evaluation.
Convert a full-scale tolerance before comparing numbers.
For this example only, suppose the controlled specification allows an error of ±1% of a 10 bar full-scale value across the stated range. The numerical limit is 10 × 0.01 = 0.100 bar. The allowed error interval is therefore -0.100 to +0.100 bar.
At a 2 bar operating point, 0.100 bar is 5% of that operating pressure. This calculation explains why a percentage of full scale and a percentage of reading produce different limits.
If the specification instead stated ±1% of reading, the limit at 2 bar would be 0.020 bar. The example's +0.030 bar rising error lies inside the first limit and outside the second. Copy the exact tolerance basis into your review record before reaching a conclusion.
Check the acceptance rule at the boundary.
ILAC G8 explains simple acceptance and decisions using guard bands. The rule determines how uncertainty affects a conformity statement and should be agreed upon before measurement where applicable. ILAC G8:09/2019, sections 4 and 5.
Consider another illustrative result with error +0.090 bar, expanded uncertainty U = 0.020 bar, and a symmetric tolerance of ±0.100 bar:
| Assumed rule | Calculation | Outcome |
|---|---|---|
| Simple acceptance: accept when absolute error is at most 0.100 bar. | 0.090 ≤ 0.100 | Accepted under this rule |
| The binary guard band of U: accept when the absolute error + U is at most 0.100 bar. | 0.090 + 0.020 = 0.110 | Not accepted under this rule |
The second rule is an example, not a universal ISO or NABL formula. It does not prove that the unknown error exceeds the tolerance. It shows that this result fails the chosen acceptance condition. Use the certificate's actual rule and uncertainty statement; do not silently substitute this one.
Record what supports return to use.
NIST explains that traceability requires a documented calibration chain with uncertainty contributions. It also states that traceability alone does not guarantee fitness for purpose. The uncertainty must suit the measurement need. NIST metrological traceability policy and FAQ.
Keep a short review record beside the certificate. The following is a suggested internal format, not a prescribed accreditation form:
- Certificate number, revision, and instrument identity.
- Intended pressure mode, operating region, and application.
- Specification reference and numerical tolerance basis.
- Reviewed results, uncertainty statement, and decision rule.
- Any missing evidence or question sent to the laboratory.
- Reviewer, date, decision, and any authorized restrictions.
If an adjustment was made, request clarification about which results describe the instrument before the change. Keep any review of previous measurements as a separate recorded action. A useful question is, which earlier jobs used the affected pressure region, and what independent evidence is available for those jobs?
If a report leaves the pressure mode, error sign, or conformity basis unclear, send a focused question identifying the certificate and result row. Record the laboratory's clarification with the review rather than guessing from the column heading.
Common questions
How is error calculated in the worked example?
Error equals gauge indication minus reference pressure. Follow the sign convention in your actual certificate.
Does 1% full scale mean 1% of every reading?
No. In the example, 1% of a 10 bar full-scale value is 0.100 bar. At a 2 bar reading, 1% of the reading would be 0.020 bar.
Can I decide acceptance from the error column alone?
You also need the applicable tolerance and decision rule. Check how that rule handles the reported uncertainty.
Can I use this table as a calibration certificate?
No. The numbers are teaching examples, not measured results or evidence for any instrument.
Prepared with AI assistance using the linked EURAMET, ILAC, and NIST publications. Calculations are illustrative. No laboratory measurements or independent technical sign-off are claimed.
Comments
Post a Comment