Pressure gauge range selection has a defined answer: divide the maximum credible pressure by 0.75 for steady service or 0.65 for pulsating service, round up to the nearest standard range, then confirm that normal operating pressure still lands in the middle third of the dial. This guide gives the formula and its source in EN 837-2, the standard range series, three fully worked examples in bar and millibar, and the ratio at which a larger range stops helping.
The pressure gauge range formula
Pressure gauge range selection is arithmetic once you have one number from the process side: the highest pressure the instrument will actually see. The working formula is minimum full-scale range = maximum credible pressure ÷ 0.75 for steady load, or ÷ 0.65 for pulsating load. Round the result up to the nearest standard range, then verify it.
- Establish normal operating pressure (Pnormal).
- Establish maximum credible pressure (Pmax): start-up, pump shut-off head, valve switching, relief valve setting.
- Decide whether the load is steady or pulsating.
- Divide Pmax by 0.75 for steady service, 0.65 for pulsating service.
- Round up to the nearest standard full-scale range.
- Check that Pnormal still falls in the middle third of the dial, roughly 33–66% of full scale.
- Check accuracy at the working point: full scale × accuracy class ÷ Pnormal.
Steps 6 and 7 are the ones most selection guides leave out. Rounding up always pushes the working point lower on the dial, so the range you order is never quite the range you calculated.
Ver faixas de manômetros no catálogo →Mais de 143 modelos de manômetros industriais→Why range is an accuracy decision, not a safety margin
Mechanical gauge accuracy is stated as a percentage of full scale (span), not of the live reading. The accuracy classes in EN 837-1 — 0.1, 0.25, 0.6, 1.0, 1.6, 2.5 and 4.0 — all refer to span, so the permitted error stays constant in engineering units right across the dial while the pressure you are measuring does not.
The consequence is direct: oversizing multiplies the error exactly where you care about it. A class 1.6 gauge reading 100 bar behaves like this.
| Working pressure | Full-scale range | Working point on dial | Permitted error, class 1.6 | Error as % of reading |
|---|---|---|---|---|
| 100 bar | 160 bar | 62.5% | ±2.56 bar | ±2.6% |
| 100 bar | 250 bar | 40.0% | ±4.00 bar | ±4.0% |
| 100 bar | 400 bar | 25.0% | ±6.40 bar | ±6.4% |
| 100 bar | 600 bar | 16.7% | ±9.60 bar | ±9.6% |
Two penalties stack up. The relative error at the working point grows in direct proportion to the oversizing, so a gauge chosen "one size up for safety" is measurably worse at its own duty point. Separately, the pointer sits low on the scale, where divisions are physically the same size but represent a larger share of the reading, so resolution drops at the same time. Oversizing costs accuracy twice, and buys protection that belongs to accessories rather than to the dial.
Where the 75% and 65% coefficients come from
EN 837-2, Pressure gauges — Part 2: Selection and installation recommendations for pressure gauges, recommends that the maximum pressure load should not exceed 75% of the full-scale value under steady load, or 65% under fluctuating load. It is a selection recommendation, not a manufacturing requirement, and it is the reason the divisor in the formula is 0.75 or 0.65.
The wording repays attention. The coefficient applies to the maximum load the gauge will see, not to normal operating pressure. Guidance that says "keep working pressure between 25% and 75% of full scale" merges two different variables: a system whose normal pressure sits at 70% of span while transients reach 95% does not meet the recommendation, even though the normal figure looks acceptable.
Manufacturer selection literature separately recommends placing normal operating pressure in the middle third of the dial, roughly 33% to 66% of span. Treat that as widely recommended practice for readability rather than as a clause you can cite, and use it as the step 6 check rather than as the sizing rule. The two solve different problems: the coefficient protects the elastic element, the middle third protects the reading.
Ver as classes de exatidão EN 837-1 em detalhe →→Standard ranges, rounding up and unit conversion
Dials are printed in production batches, so ranges come from a decimal series rather than from your calculation. The series in common catalogue use runs 1, 1.6, 2.5, 4 and 6 in each decade: 1, 1.6, 2.5, 4, 6, 10, 16, 25, 40, 60, 100, 160, 250, 400, 600 and 1000 bar. A calculated minimum of 267 bar becomes an ordered range of 400 bar.
That rounding step is not free. Going from 267 to 400 bar drops the working point further down the dial and raises the error there in the same proportion, which is why step 7 must be repeated on the range you will order rather than on the number you calculated. Where the jump is large, a tighter accuracy class is usually the cheaper correction.
Ranges are quoted in different units by region, so convert before comparing catalogues against a specification.
| 1 bar equals | Value |
|---|---|
| psi | 14.5038 |
| MPa | 0.1 |
| kPa | 100 |
| kgf/cm² | 1.0197 |
Worked example 1: pump discharge, steady load
A centrifugal pump delivers 8 bar in normal operation. The pump curve gives a shut-off head equivalent to 12 bar, the highest pressure the discharge gauge can credibly see, and the service is steady.
Dividing by the steady-load coefficient gives 12 ÷ 0.75 = 16.0 bar, already a standard range, so no rounding is needed. Normal pressure sits at 8 ÷ 16 = 50% of span, comfortably in the middle third. At class 1.6 the permitted error is 16 × 1.6% = ±0.26 bar, or ±3.2% of the 8 bar reading.
This is the shape of a calculation that works: the maximum is covered, the pointer sits near mid-scale, and the error at the duty point is acceptable for pump monitoring. If the duty needed better than ±2% at the working point, the correct move is a tighter accuracy class on the same 16 bar range, not a different range. For placement and connection detail see the centrifugal pump gauge selection guide.
Worked example 2: hydraulic power unit, pulsating load
A hydraulic power unit runs at 100 bar with the relief valve set at 140 bar. Pump ripple makes this a pulsating duty, and switching spikes reach roughly 250 bar, which a snubber damps before they reach the instrument.
Sizing on the relief setting with the pulsating coefficient gives 140 ÷ 0.65 = 215.4 bar, rounding up to a standard 250 bar range. Normal pressure lands at 100 ÷ 250 = 40% of span, and class 1.6 gives ±4.0 bar, or ±4.0% of the 100 bar reading.
The instructive part is the alternative. Sizing on the undamped 250 bar spike gives 250 ÷ 0.65 = 384.6 bar, rounding up to 400 bar. That range pushes the working point down to 25% of span and the error at 100 bar up to ±6.4%. Buying a bigger dial to absorb spikes costs 60% more error at the duty point than fitting a snubber and sizing on the relief setting. Circuit placement is covered in the hydraulic gauge selection guide.
Worked example 3: low-pressure nitrogen blanketing, where the sum does not close
A nitrogen blanketing header holds 20 mbar in normal operation, and the regulator can credibly deliver 50 mbar on a fault. The load is steady, so 50 ÷ 0.75 = 66.7 mbar, which rounds up to a 100 mbar capsule range. Normal pressure then sits at 20 ÷ 100 = 20% of span: the middle-third check fails.
Dropping to the next standard range does not rescue it. A 60 mbar range would put the working point at 33%, but 60 × 0.75 = 45 mbar is below the 50 mbar maximum, so it breaks the sizing recommendation instead. No standard range satisfies both conditions, because the ratio of maximum to normal pressure is 50 ÷ 20 = 2.5.
The calculation is telling you the range is not the variable to adjust. The options are a pressure limiter that caps the maximum so a 60 mbar range becomes valid, or accepting the low-scale reading and recovering measurement quality with a tighter accuracy class.
| Step | Pump discharge | Hydraulic power unit | Nitrogen blanketing |
|---|---|---|---|
| Pnormal | 8 bar | 100 bar | 20 mbar |
| Pmax and source | 12 bar, shut-off head | 140 bar, relief setting | 50 mbar, regulator fault |
| Load type and coefficient | Steady, 0.75 | Pulsating, 0.65 | Steady, 0.75 |
| Calculated minimum | 16.0 bar | 215.4 bar | 66.7 mbar |
| Standard range ordered | 16 bar | 250 bar | 100 mbar |
| Working point on dial | 50% ✓ | 40% ✓ | 20% × |
| Error at working point, class 1.6 | ±3.2% | ±4.0% | ±8.0% |
When the calculation says a bigger gauge will not help
The third example generalises. Before rounding, the working point equals the coefficient multiplied by Pnormal ÷ Pmax. Requiring that to reach the 33% readability floor limits how far apart the two pressures can be: Pmax ÷ Pnormal must not exceed about 2.3 for steady service (0.75 ÷ 0.33), or about 2.0 for pulsating service (0.65 ÷ 0.33). Rounding up lowers the working point further, so this is a necessary condition, not a sufficient one.
Above that ratio no range satisfies both rules, and the answer is to reduce Pmax rather than raise the range: a snubber for pulsation, a pressure limiter for excursions, or a digital gauge with peak hold where the transient itself must be captured. Overpressure protection is an accessory decision, not a dial decision.
Two figures that look contradictory are worth separating. EN 837-1 states pressure limits for Bourdon tube gauges as steady load at the full-scale value, fluctuating load at 0.9 times full scale and short-time load at 1.3 times full scale, while EN 837-2 recommends keeping the maximum load below 75%. The first describes what the instrument withstands; the second is what to select for service life, zero stability and readability. The standard also distinguishes full-scale loadable from non-full-scale loadable designs, so confirm the limit for the specific model against its data sheet.
What range calculation cannot decide
This procedure sizes a dial. It does not confirm the inputs. Pmax, the medium, the temperature and whether the load is genuinely steady all come from the process side, and a range calculated from an optimistic maximum is wrong however carefully it is rounded.
Getting the range right also does not complete the selection. Wetted material, process connection, filling, ingress protection, dial size and any hazardous-area, oxygen or hygienic requirements are separate decisions, and a correctly ranged gauge in the wrong material still fails.
Relief valve settings and design pressures belong to process and equipment engineering: gauge selection reads those numbers, it never changes them. For safety-related service, have the responsible engineer confirm the range, accessory setpoints and materials before release.
Key takeaways
- Size the range from maximum credible pressure divided by 0.75 for steady load or 0.65 for pulsating load, then round up to a standard range.
- Accuracy is a percentage of span, so an oversized range multiplies the error at the working point in direct proportion.
- If maximum pressure exceeds normal pressure by more than about 2.3 times (2.0 pulsating), no range satisfies both rules and the answer is an accessory.
- The calculation sizes a dial only: P_max, medium, temperature and safety-related settings must be confirmed by the responsible process engineer.
Perguntas frequentes
Is a bigger pressure gauge range safer?
No. Accuracy is a percentage of full scale, so oversizing multiplies the error at your working point. A class 1.6 gauge reading 100 bar gives ±2.6% of reading on a 160 bar range but ±6.4% on a 400 bar range. Spike protection belongs to snubbers and limiters, not to a larger dial.
Why is a 400 bar gauge inaccurate at 100 bar?
Class 1.6 means ±1.6% of the 400 bar span, which is ±6.4 bar everywhere on the dial. At a 100 bar reading that fixed ±6.4 bar is ±6.4% of the value read, while the same absolute error at 250 bar would be only ±2.6%.
Is the 75% rule a standard requirement?
A recommendation, not a requirement. EN 837-2, the selection and installation part of the standard, recommends that maximum pressure load stays below 75% of full scale under steady load and 65% under fluctuating load. It applies to the maximum load, not to normal operating pressure.
What if the calculated range is not a standard range?
Round up to the next range in the 1, 1.6, 2.5, 4, 6 decimal series, then repeat the verification on the range you will order. Rounding up lowers the working point and raises the error there in the same proportion.
How do I calculate the range for a compound or vacuum gauge?
Use the same coefficients, but base the working point check on the whole span. A -1 to 0 to 15 bar dial spans 16 bar, so a 5 bar working pressure sits at 6 bar of span, or 37.5%, not at a third of the positive section.
How do I convert a bar range to psi when comparing catalogues?
Multiply by 14.5038. A 16 bar range is 232 psi and 250 bar is 3626 psi, which is why catalogue psi ranges look irregular: they are conversions of the bar series. 1 bar also equals 0.1 MPa, 100 kPa and 1.0197 kgf/cm².