Beck publishes no fixed recalibration interval for the 981–990 transmitters, so you derive it. Start from the datasheet drift figure of ±0.3% FS per year, divide your tolerance budget by it, and check the loop at that interval. A field check is two points: zero with both ports at atmosphere, then span against a reference.
Everything below uses figures from the Beck datasheets, so the numbers can be checked against the source. The one thing no datasheet will give you is your own tolerance requirement, and that is the number the whole interval hangs on.
What the datasheet actually promises
All Beck figures are quoted at the reference conditions of EN 60770, which means 23 °C. Read them with that in mind, because a field reading taken at 5 °C or 45 °C will not reproduce them.
On the 988, the measuring cell accuracy is ±0.4% FS typical and ±1.0% FS maximum, both including non-linearity, hysteresis and repeatability. Hysteresis is separately limited to ±0.2% FS and repeatability to 0.1% FS. Long-term drift is limited to ±0.3% FS per year. Total error over −20 °C to +85 °C is quoted at ±1.0% FS for the 3-wire versions and ±1.5% for the 2-wire version, and the exact figure for your version should be read from its own datasheet.
The 985 quotes linearity, again including hysteresis and repeatability, at ±0.5% FS with a minimum of ±1 Pa, and a total error band excluding long-term and temperature effects of ±1% FS. The split matters: the ±1% figure deliberately leaves drift and temperature out, so adding them back is your job.
Two more terms are worth knowing because they behave differently from drift. Temperature effect on the 988 is 0.018% per K typical and 0.036% per K maximum. Supply voltage effect is under 0.04% per V. Neither accumulates with time, so they do not shorten the interval, but they do explain a reading that looks wrong on the day you check it.
Setting the interval: work backwards from your budget
Take the allowance your process can live with, in percent of full scale, and divide it by 0.3% FS per year.
| Allowance (FS) | Time for drift to reach it | Practical interval |
|---|---|---|
| 0.2% | 0.67 yr (about 8 months) | 6 months |
| 0.3% | 1.0 yr | 12 months |
| 0.5% | 1.7 yr (about 20 months) | 12 months, trend between checks |
| 1.0% | 3.3 yr (about 40 months) | 24 months |
The table assumes the transmitter sits near the reference temperature. It usually does not. A transmitter in a plant room that swings 40 K across the seasons picks up about 0.72% FS from the thermal term alone at the typical figure, and up to 1.44% FS at the maximum. That single term can consume an entire 1% allowance, so where the temperature swing is the dominant error, shorten the interval rather than trusting the drift arithmetic.
The other reason to check more often than the arithmetic suggests is that drift is not always drift. A loop losing 0.3% per year because the piezoresistive cell is ageing is normal. A loop losing 0.3% per quarter usually has a wet impulse line, a partially blocked port or a supply problem. Record the as-found error at every check, not just the correction you applied, and the pattern tells you which one you have.
The field check: two points, and only two
- Isolate the transmitter and let both ports sit at atmosphere. Read the output. On a 4–20 mA device that is 4 mA, on a 0–10 V device 0 V. Wait for the reading to settle, because the 988 quotes a response time of 100 to 200 ms and a measurement cycle of 2 to 4 ms, and both are faster than a warm-up settling.
- Apply a known reference pressure near full scale. Read the output again.
- If zero is out, correct it using the means your version provides. On the 985, version M is manual offset compensation and A is automatic; Q is automatic with eight selectable ranges. On the 990, 990M gives manual zero and 990A automatic zero. The 985Q selects ranges with a rotary switch and the 985M and 985A with a jumper.
- Re-check both points. Zero and span interact on a real device, so correcting one moves the other. Two passes are normal.
- Write down the as-found and as-left values at both points, with the date and the reference you used.
Three things go wrong in the field more often than the transmitter does. A differential transmitter zeroed with water sitting in an air line reads that water as pressure, so the “correction” you just made is wrong by exactly the amount of water. A multi-range device zeroed before confirming which range is selected will be calibrated against the wrong span. And a 2-wire 4–20 mA loop needs a load within the datasheet limit, ≤400 Ω at 24 VDC on the 988, or the reading you are adjusting is not the reading the controller sees.
Do not expect a field check to reproduce a bench calibration. EN 60770 reference conditions are 23 °C, and nothing you do on site changes the temperature of the cell.
What actually causes the drift
Ageing of the piezoresistive cell is the only cause the datasheet covers, at up to ±0.3% FS per year. Everything else that shifts a reading is either recoverable or avoidable.
Temperature is the biggest recoverable term, at 0.018% per K typical. If a transmitter is mounted where it sees the medium temperature rather than ambient, remember the 988 medium rating runs to +105 °C for short periods and the ambient rating ends at +85 °C, so the electronics can be well outside the reference conditions while the process is still inside its own limits.
A sagging supply costs under 0.04% per V, so a 3 V drop on a nominal 24 V supply moves the output by about 0.12% FS. Pressure cycling produces hysteresis of up to ±0.2% FS and repeatability of 0.1% FS, which shows up as a difference between the rising and falling readings at the same point.
Then there is the installation. Clogged or kinked impulse lines, condensation in the air lines of a differential transmitter, a diaphragm contaminated by the medium, a mounting bracket tightened hard enough to stress the housing, and over-range events all shift the reading without any ageing of the cell at all. Check these before you conclude the transmitter has drifted.
One selection error is worth naming, because it is not a calibration problem and cannot be calibrated out. The 988 is built for liquid circuits with a ceramic Al₂O₃ cell and Parylene C coating. The 985 and 990 are specified for air and other non-combustible, non-aggressive gases. Fitting a gas transmitter on a liquid line, or a liquid transmitter where condensation will form in the ports, produces exactly the symptoms of drift.
Which Beck transmitter, and what it means for calibration
The series differ in what you can adjust and how, so the procedure is not identical across the range.
| Series | Medium | Span | Output | Range selection |
|---|---|---|---|---|
| 981 | liquids and gases | ±50 mbar to 0…600 mbar | 0–10 V (3-wire) or 4–20 mA (2-wire) | fixed |
| 982R | air and non-aggressive gases | 0…100 Pa to 0…10 kPa | analogue | single adjustable range, one-button 50% reduction |
| 984 | differential, overpressure, vacuum | multiple | two outputs, by jumper or rotary switch | up to 8 switchable ranges |
| 985 | differential, overpressure, vacuum | −100 Pa to 1 bar | 0–10 V, 4–20 mA or Modbus RTU | 985Q eight ranges by rotary switch, 985M/985A by jumper |
| 988 | liquids | 0…2 / 0…6 / 0…10 bar | 0–10 V or 4–20 mA | fixed |
| 990 | air and non-aggressive gases | 0 Pa to 2,500 mbar | 0–10 V or 4–20 mA, menu-selectable | fixed, with unit switching and square-root extraction |
The practical consequence is that a filter-monitoring loop built on 985Q devices can be re-ranged and zeroed on site without opening the enclosure, while a 988 in a cooling loop is set at the factory and a field check is verification rather than adjustment. If your maintenance plan assumes every transmitter can be re-ranged in place, that assumption holds for the 984 and 985 families and not for the 988.
What to send us for a quotation
Tell us the medium, the span you need, the output your controller expects, and whether the transmitter has to survive vibration, a hazardous area or washdown. With those four we can name the series and version, and flag the ones that can be adjusted on site against the ones that cannot. We supply the Beck range together with the matching connection and mounting hardware.
- Browse the Beck Sensortechnik range we carry.
- See the wider instrument and sensor category.
- Compare the full product range we supply.
- Read more selection and maintenance guides in the Insights library.
- Send your requirements straight to the inquiry form.
Every Beck device we supply is original and new, bought through the manufacturer or established distribution channels, with procurement documentation for verification on arrival.