Two facts decide a defensible calibration routine: S+S ships every sensor factory-calibrated, and the datasheets state long-term stability of ±1 percent per year. With a typical storage alarm window of 5 percent RH, annual recalibration is enough. If your window is 3 percent RH, verify every six months and keep the as-found record. Everything else is procedure.
Figures come from S+S Regeltechnik datasheets for the HYGRASGARD AFTF-Modbus-T3, AFTF-35 and KFTF-35, plus the S+S article on humidity in pharmaceutical logistics.
Why readings move in pharma storage and cold rooms
A digital humidity sensor does not fail suddenly. It walks away from the truth, faster where you need it most.
- Condensation. Warm air crossing into a cold room settles water on the element. Readings turn sluggish and the control signal stops representing the room.
- Cleaning agents. Alcohols and peroxides are volatile and hygroscopic, so they change the vapour pressure at the filter surface.
- Dust and packaging fibres. A loaded sinter filter slows the exchange of water vapour between the air and the element.
- Door openings and cargo handling. Air exchange produces humidity spikes that a controller may act on even when the sensor is fine.
None of this raises an error flag. It shows up as a quiet offset, so the routine must be scheduled, not reactive.
Calibration, adjustment and verification are three jobs
Mixing the three words is the fastest way to fail an audit.
- Calibration compares the transmitter against a reference and records the difference. It changes nothing.
- Adjustment changes the transmitter so the reading moves closer to the reference. It is only defensible after the as-found value is recorded.
- Verification is your own regular check that the transmitter still sits inside the tolerance you declared.
A certificate with no as-found value cannot tell you the direction or the size of the drift.
Setting an interval you can defend
Start from published drift, not from habit. S+S states ±1 percent per year for the AFTF-35 and KFTF-35.
Then divide by your own tolerance. A room alarm at 60 percent RH with a 5 percent RH action window gives one year of drift a fifth of the budget, so annual calibration is comfortable. A product that must hold 45 to 50 percent RH gives a 3 percent RH window, and one year of drift eats two thirds of it. Verify every six months and calibrate annually.
Three inputs sharpen it further: your own drift history after two cycles; the consequence of a wrong reading, where a value that releases a batch deserves a shorter interval than one that only triggers a report; and the measurement point, since a door-side or airlock sensor sees more condensation cycles than an internal room sensor.
An interval with a stated reason survives an inspection. One that copies a supplier recommendation does not.
Which calibration document you actually need
Keep two levels running at once: in-house verification for the routine, and an accredited certificate for the points that release product or appear in a validation file. A factory certificate is a shipping document, not audit evidence.
Field verification that survives an audit
This assumes you hold a calibrated reference, such as a chilled-mirror hygrometer.
- Record the as-found value first. Read the transmitter and the reference at the same time, at the same height, at least 300 mm apart and away from direct airflow.
- Respect the timing figures. The AFTF-35 and KFTF-35 give t90 below 60 seconds and warm-up below 10 minutes, so a reading taken two minutes after power-up is not a measurement. Allow minutes, and longer in a cold room.
- Avoid disturbed conditions. Do not verify straight after a door cycle, a cleaning shift or a defrost.
- Inspect the filter before you judge the electronics. Both the AFTF-Modbus-T3 (plastic sinter, 13 mm) and the 35 series (16 mm, 35 mm long, optional metal version) use exchangeable filters. A contaminated filter mimics drift and no adjustment fixes it.
- Check the position against the storage geometry. S+S recommends good air movement, not directly at doors, and not in the airflow from refrigeration units, vents or fans.
- Record the as-left value, the reference and its due date, and who signed it off.
Use one template at every point, and keep the failed checks in the file. A file with only passes reads as incomplete.
What you can adjust
The AFTF-Modbus-T3 datasheet gives a zero point offset range of ±10 percent RH for humidity and ±5 degrees C for temperature, with further offset correction over Modbus, and states that the sensor is factory-calibrated with expert environmental precision adjustment possible.
That is useful and easy to misuse.
- The offset is a trim, not a repair. A 10 percent RH trim aimed at a sensor that has drifted 15 percent RH leaves you further from the truth with a worse record.
- Adjust at your operating point, not in a workshop at a different humidity.
- Re-verify after adjusting. An adjustment without a following verification is an unproven change.
- Know where the offset lives. On the Modbus version the raw and corrected values are both available over the bus. On the analogue variants the trim is a local setting that needs a written note.
For a computerised system under GxP the Modbus route is easier to defend: the trend before and after an adjustment can be exported rather than reconstructed from paper.
Which S+S transmitter makes the routine easiest
Three choices follow, rather than a list:
- Where condensation forms, including cold rooms, airlocks and loading areas: AFTF-35 or KFTF-35. The heating function and the higher permitted humidity are the deciding features, and the ±3.0 percent class is the price.
- Conditioned storage and clean rooms held between 20 and 80 percent RH: AFTF-Modbus-T3. It carries the ±2.0 percent class where that applies, and the bus allows drift trending without entering the room.
- Duct measurement for air handling: KFTF-35, whose PLEUROFORM probe is rated for air velocities up to 30 m/s.
Where cabling a cold room is the expensive part, W-Modbus replaces the RTU cable with a 2.4 GHz link, quoted at up to 500 m in open field and 50 to 70 m inside buildings. Treat that as best case.
Where to source it
S+S Regeltechnik manufactures in Nuremberg, tests and calibrates every device before delivery, and gives a five year guarantee on new products. The factory state is the first entry in every drift history.
We supply S+S Regeltechnik devices as a distributor of genuine, brand-new originals, including the AFTF on-wall, KFTF duct, RFTM room and HFTM humidity families. Start at the brand page, narrow to the temperature and humidity slice, or go straight to the AFTF wall and duct sensors, the KFTF duct series and the RFTM room units. Worked example: our pharma clean room case; control strategy in controlled environment solutions; selection criteria in our selection article. For mixed enquiries use the brand index, the full product catalogue or the other application cases, or send the measurement point list through the contact form.
FAQ
How often should S+S humidity transmitters be calibrated?
Start at annual and justify any change from your own data. With ±1 percent per year of stated stability, a 5 percent RH alarm window can stay annual, while a 3 percent RH window should be verified every six months. After two cycles, use your measured drift, not the datasheet.
Can I calibrate an S+S transmitter on site, or must it go to a laboratory?
Verify it on site against a calibrated reference; that is the working routine in most plants. A formal traceable calibration with stated uncertainty needs an accredited laboratory, because that is what the certificate represents. The AFTF-Modbus-T3 datasheet notes the sensor is factory-calibrated and that expert environmental precision adjustment remains possible.
Does a factory calibration certificate satisfy a GMP audit?
Usually not on its own. It is traceable to the manufacturer's own references rather than to national standards through an accredited scope, so it belongs in incoming goods records. For points that release product, buy a certificate to DIN EN ISO/IEC 17025 from an accredited laboratory and keep the factory certificate as the baseline entry.
Why does my transmitter read high after a cleaning shift?
Most often the filter is wet or loaded, not the electronics. A wet sinter filter holds a local microclimate, so the reading lags and then overshoots. Let the element settle, respect the stated response time, and inspect the filter. Both filters here are exchangeable, and replacing one is cheaper than an adjustment.
Which accuracy figure should I write into the procedure?
Write the figure with its condition, or it will be read as a blanket tolerance. For the AFTF-Modbus-T3 that is ±2.0 percent between 20 and 80 percent RH at +25 degrees C, and ±3.0 percent outside. For the 35 series it is ±3.0 percent between 30 and 70 percent RH at +25 degrees C, and ±3.5 percent outside. A point at 85 percent RH is held to the wider figure.