Instrumentation basics · 1.4
Range, span, accuracy, repeatability and turndown, with the numbers
25 October 2026 · 10 min read · IEC 60770-1, IEC 61298-2
In short
- The sensor limits (LRL/URL) are what the instrument can do. The calibrated range (LRV/URV) is what you set it to. Span = URV − LRV.
- Turndown = URL ÷ span. The higher the turndown, the larger the error as a share of what you are measuring.
- Always ask "% of what?" An error of 0.5 % of span is five times larger than 0.5 % of reading at 20 % of range.
- Real accuracy is the total probable error: reference accuracy combined with temperature, static pressure and stability effects.
Every datasheet has a range and an accuracy line, and both are easy to fill in wrongly. This article defines the terms the way vendors use them, then works through one transmitter specification to show what a "0.075 % accurate" instrument really delivers in the field.
Where this matters
- Trips and alarms: if the error is larger than the gap between the normal value and the trip point, the plant trips when nothing is wrong, or fails to trip when it should.
- Custody transfer and fiscal metering: an error of a fraction of a percent on a large flow is real money every day.
- Control quality: a noisy or poorly repeatable measurement limits how tight a loop can be.
- Spares and standardisation: picking one sensor size to cover many ranges is cheaper, but only if the turndown stays reasonable.
How it works
Limits, range and span
| Term | Meaning | Example |
|---|---|---|
| URL / LRL | Upper and lower range limit: the most the sensor can be set to | 0–70 bar |
| URV / LRV | Upper and lower range value: the calibrated 20 mA and 4 mA points | 0–10 bar |
| Span | URV − LRV | 10 bar |
| Turndown | URL ÷ span | 70 ÷ 10 = 7 : 1 |
| Maximum turndown | URL ÷ minimum allowed span (vendor limit) | often 100 : 1 or more |
A suppressed zero (LRV above zero, such as 5–10 bar) and an elevated zero (LRV below zero, such as −2 to +3 m in DP level) are both normal. They are common in level measurement with wet legs (article 5.1).
Accuracy terms
- Accuracy (or measured error) is how close the reading is to the true value. It includes non-linearity, hysteresis and repeatability at reference conditions.
- Repeatability is how close repeated readings of the same value are, approached from the same direction. An instrument can be repeatable but inaccurate, which calibration can fix.
- Hysteresis is the difference between readings taken going up and coming down.
- Stability (drift) is the change over time, quoted as % of URL per year or over five or ten years.
- Resolution is the smallest change the instrument shows.
"% of what?"
Pressure and temperature transmitters usually quote % of span. Magnetic and Coriolis flowmeters usually quote % of reading. Gauges quote an accuracy class as % of full scale. The difference is large at low readings. An error of ±0.5 % of span at 20 % of range is ±0.5 / 0.2 = ±2.5 % of reading.
Total probable error
The accuracy on the front of a brochure is at reference conditions: about 20 °C, steady and freshly calibrated. In the field, other effects add to it. Because they are independent, they are combined by root-sum-square:
Eref is the reference accuracy, Etemp the ambient temperature effect over the expected change, and Estatic the static (line) pressure effect, which applies to DP transmitters. Stability is assessed separately against the calibration interval.
What the standards and practice say
- IEC 60770-1 / -2 sets how transmitter performance is evaluated and stated. IEC 61298-2 sets the test methods behind datasheet figures.
- Vendors quote reference accuracy at a stated turndown. Beyond it, many add a term that grows with turndown, so check the footnotes.
- A common rule of thumb is to keep turndown at or below 10 : 1 for critical measurements, unless the vendor's specification shows the error is still acceptable.
- The ranges on the datasheet should cover normal operation at roughly 50–75 % of span, with headroom for upsets. Trip points must be inside the range with margin.
Worked example: what "0.075 %" really means
Here is an example gauge pressure transmitter specification, of the kind found on vendor datasheets:
| Specification item | Example value |
|---|---|
| Sensor URL | 70 bar |
| Reference accuracy | ±0.075 % of span |
| Ambient temperature effect | ±(0.025 % URL + 0.125 % span) per 28 °C |
| Expected ambient change from calibration | 20 °C |
Case 1, ranged 0–10 bar (turndown 7 : 1).
- Eref = 0.075 % × 10 = 0.0075 bar
- Etemp = (0.025 % × 70 + 0.125 % × 10) × 20/28 = (0.0175 + 0.0125) × 0.714 = 0.0214 bar
- TPE = √(0.0075² + 0.0214²) = 0.0227 bar = 0.23 % of span
Case 2, same sensor ranged 0–2 bar (turndown 35 : 1).
- Eref = 0.075 % × 2 = 0.0015 bar
- Etemp = (0.0175 + 0.0025) × 0.714 = 0.0143 bar
- TPE = √(0.0015² + 0.0143²) = 0.0144 bar = 0.72 % of span
| Range | Turndown | Brochure accuracy | Total probable error |
|---|---|---|---|
| 0–10 bar | 7 : 1 | 0.075 % | 0.23 % of span |
| 0–2 bar | 35 : 1 | 0.075 % | 0.72 % of span |
What this example shows
- The brochure figure is the smallest term. Here the temperature effect is three to ten times larger than the reference accuracy.
- The URL term does not shrink with the range. At high turndown it dominates, which is why the error as a share of span triples.
- The fix is a smaller sensor. A sensor with URL 7 bar, ranged 0–2 bar (turndown 3.5 : 1), would take the URL term down tenfold.
When it fails: the fixes in order
- Pick a sensor whose URL suits the range. This is the cheapest accuracy improvement there is.
- Reduce the ambient change. Sunshades, enclosures or calibrating near the operating temperature all help.
- Specify the accuracy class needed. Premium "0.04 %" transmitters exist, but they help only if the other terms are also small.
- Shorten the calibration interval where stability is the limit.
- For flow, check the % of reading at minimum flow, not just at full scale.
Common mistakes
- Comparing a % of span figure with a % of reading figure as if they were the same.
- Using the brochure accuracy for a trip margin, and ignoring the temperature and static pressure effects.
- Standardising on one high-URL sensor and ranging it far down.
- Setting a range so that the normal value sits at 10 % or 95 % of span.
- Putting a trip point outside the calibrated range, or right at its end.
- Confusing repeatability with accuracy, for example accepting a repeatable reading that was never calibrated.
Checklist
- LRV, URV and units are stated, and the URL of the chosen sensor is known.
- Turndown is calculated and within the vendor's limit for the stated accuracy.
- The normal value sits at roughly 50–75 % of span.
- Alarm and trip points are inside the range with margin.
- The accuracy basis (% of span, reading or URL) is stated.
- The total probable error is estimated for the expected ambient change, and static pressure for DP.
- The error is smaller than the gap between normal operation and the trip.
- Stability is checked against the planned calibration interval.
Quick reference
- Span = URV − LRV. Turndown = URL ÷ span.
- % of span → % of reading: divide by the fraction of range (e.g. 0.5 % ÷ 0.2 = 2.5 %).
- TPE = √(Eref² + Etemp² + Estatic²)
- Rule of thumb: turndown ≤ 10 : 1 for critical measurements, normal value at 50–75 % of span.
Further reading
- IEC 60770-1, Transmitters for use in industrial-process control systems: methods for performance evaluation
- IEC 61298-2, Process measurement and control devices: general methods and procedures for evaluating performance (tests under reference conditions)
- In this series: 1.2 The instrument loop · next, 1.5 Gauge, absolute and differential pressure
The transmitter specification is an illustrative example in the usual vendor form, not a particular product. Use the actual vendor's specification, including its footnotes on turndown, for design. IEC is named for identification only.