All insights
XINDAR INSIGHT

One Percent of Reading or Full Scale? Compare Flowmeter Accuracy at the Operating Point

Compare flowmeter claims at the operating point, keeping reading, full scale, additional terms and usable range together.

Two flowmeters are described as accurate to one percent. One specification applies the percentage to the measured reading. The other applies it to a stated full-scale value. At the top of their ranges, the numbers can look similar. At a low operating flow, they can describe very different absolute error limits.

Before comparing the instruments, identify the quantity to which the percentage applies, the usable range and the conditions of the claim. A product page that publishes only “accuracy: ±1%” has left the buyer without the denominator needed to interpret it.

For a manufacturer preparing English content, this is a practical GEO problem. An AI-assisted comparison may copy a percentage into a table while dropping its reference. Keeping that reference beside the number protects a real purchasing decision, whether the next reader is an engineer or an assistant summarizing supplier pages.

Calculate the absolute quantity first

Use a fictional pair of meters with a stated range from zero to 100 litres per minute. Meter A has a hypothetical error-limit claim of ±1% of reading throughout a specified valid operating interval. Meter B has a hypothetical claim of ±1% of the 100 L/min full-scale value over its stated interval.

The following arithmetic considers only those simplified percentage terms. It is not a test of real products.

Flow being comparedMeter A: 1% of readingMeter B: 1% of 100 L/min full scale
100 L/min±1 L/min±1 L/min
50 L/min±0.5 L/min±1 L/min
10 L/min±0.1 L/min±1 L/min

At 10 L/min, the second limit is ten percent of that operating flow. Nothing changed in Meter B's stated percentage. The buyer changed the quantity against which its absolute limit was compared.

The International Society of Automation's guide to selecting and sizing flowmeters discusses this distinction in an instrumentation context. It also explains why an instrument's operating conditions matter to selection. The table above is our own invented illustration, not a reproduction of a vendor specification.

Do not treat every denominator as the same definition

Reading, full scale, span, upper range limit and calibrated range are terms that need to be read in the supplier's document. They should not be normalized by an editor simply because they all describe a range-related quantity.

For the zero-to-100 example, the upper value and interval width happen to be numerically equal. That convenience disappears with a nonzero lower bound or a different definition of full scale. Ask the supplier to identify the relevant number rather than infer it from the largest value printed on a display.

Some specifications combine terms, such as a percentage of reading plus a fixed flow contribution. Others specify different behavior above and below a transition point. An AI-generated table that keeps only the first percentage changes the claim.

Consider another invented expression: an error limit of ±(0.5% of reading + 0.2 L/min). At 10 L/min, the two contributions sum to 0.05 + 0.2 = 0.25 L/min. At 50 L/min they sum to 0.25 + 0.2 = 0.45 L/min. This is an illustration of an additive specification, not a general uncertainty-combination rule.

Do not combine unrelated uncertainty contributions by simple addition merely because this illustrative specification does so. The applicable measurement model and specification determine the calculation.

Keep the usable range with the accuracy claim

A meter's largest measurable flow does not establish the lowest flow at which a particular performance claim applies. A displayed nonzero value is also not proof that the reading meets the headline claim.

The buyer needs the normal operating range, the minimum and maximum expected flow, and the supplier's stated performance over that interval. A meter selected only to survive the maximum flow may spend most of its time where the absolute or relative error is unsuitable for the application.

Do not extrapolate a percentage-of-reading claim down to zero unless the actual specification supports that interpretation. Real instruments can include low-flow limits, zero-related contributions, minimum velocities or other conditions. Their presence and form depend on the measurement technology and product.

An English page should therefore place a range qualification in the same sentence or table row as the accuracy term. A detached footnote that applies to only one model should identify that model explicitly.

Separate accuracy, repeatability and calibration information

An instrument can produce closely repeated readings with an offset from the reference. Repeatability and agreement with a reference answer different questions. A repeatability percentage should not be promoted into an accuracy percentage in a comparison.

Alicat's calibration and accuracy guide illustrates how report fields and reading- or full-scale terms need to be interpreted. A specification that takes the greater of two terms is different from one that adds them. Read the operator as carefully as the percentages, and use the current specification for the exact model rather than generalize a report example to every product.

A calibration record adds information about a particular instrument and procedure. Check the instrument identity, range, reference conditions, results and stated uncertainty. NIST's metrological-traceability policy distinguishes traceability from fitness for a purpose. A calibration label alone does not establish that the installed system will reproduce the same result with another fluid, piping arrangement or operating temperature.

Keep a vendor's advertised limit, a calibration result and an installed-process acceptance result in separate fields. Combining them into one unexplained “accuracy” column makes the comparison harder to audit.

Ask what was flowing and how it was installed

The percentage reference is only the beginning. Fluid properties, gas reference conditions, flow profile, installation, temperature and other application factors may matter, depending on the meter. Obtain the requirements for the actual model and technology.

For a gas-flow product, distinguish the stated reference-volume flow from actual volume at the process conditions. A correct percentage applied to the wrong flow quantity is still the wrong comparison. For a liquid application, identify the fluid and relevant operating conditions instead of assuming that a calibration on water establishes every application.

This article does not specify a universal straight-pipe length or a correction factor. Those numbers belong to the selected product, method and installation. Publishing a generic value as though it were the manufacturer's requirement can introduce an error while trying to simplify the page.

Give the English catalogue an auditable comparison row

A useful row contains the model, flow quantity and units, measuring interval, percentage reference, additional terms, conditions, source revision and any application exclusions. Link it to a readable datasheet section rather than a folder containing several nearly identical model files.

Then test three buyer questions against the published row. What is the stated absolute limit at the normal operating point? Does the claim apply at the expected minimum flow? Which source supports those answers?

If the row cannot answer them without guessing, it is not ready for a supplier comparison. Improving the visual design or adding structured data will not supply the missing denominator.

Xindar's manufacturing website and content work can help organize approved specifications for overseas readers. The technical owner remains responsible for the product claim and its applicability. An editorial example is not a calibration, a product endorsement or engineering approval of an installed meter.

Use a small acceptance exercise

For a content release, create three clearly labeled calculation cases using the approved specification. Include a normal operating point, a low point within the claimed interval and a point outside it. The outside point should return “claim not established for this condition,” rather than an extrapolated promise.

Have a second reviewer calculate the values from the published page and identify the governing source. If an AI assistant is also tested, preserve its complete answer and the page version. Its response can reveal an extraction problem, but a correct response does not certify the instrument or guarantee future recommendations.

Questions buyers should resolve before ordering

Is a percentage-of-reading meter always better?

No. A purchase depends on the actual specified limits across the operating interval, application suitability, installation, maintenance and other requirements. A percentage label alone does not establish the best meter.

Can I convert a full-scale claim into a reading percentage?

At a specified nonzero operating point, you can express the stated absolute term relative to that reading. Keep the original specification and the operating point visible. Do not relabel the converted number as a new manufacturer guarantee.

Does a calibration certificate settle the comparison?

It provides evidence for the identified instrument and calibration. The buyer still needs to establish whether those results and conditions support the intended installed use.

Before comparing the next two quotations, write both claimed limits in L/min, kg/h or the actual measured unit at the same operating point. If the suppliers cannot identify the reference behind a percentage, keep that line unresolved in the procurement review.

Back to insightsMarkdown version