Humidity sensor reads low: why it drifts in damp air

A humidity sensor that steps down tens of points and never climbs back is contaminated on its sensing film, not miscalibrated and not reading dry air. The swap test settles it in a day: a fresh part reading correctly beside the suspect one puts the fault on the polymer. Bake the suspect part at 100 to 105 °C [212 to 221 °F] at under 5% RH for 10 h, as the sensor vendor's handling instructions describe, then re-hydrate it and keep it only if it lands within about 5 points of a trusted meter. Mount the next one behind a pressure-equalization membrane vent where the local RH stays under about 80%, because contamination, not wear, is what ends these parts early.

Why does a humidity sensor step down 30 points and stay there?

A step of about 30 points RH that never recovers is contamination of the sensing polymer: the offset is ten times the ±3% RH accuracy a BME280-class part is specified to, so the reading tracks the film, not the room.

The element is a polymer whose dielectric constant shifts as it absorbs water vapor; the chip reads that capacitance as humidity. Other polar molecules absorb into the same film and shift the same capacitance, so one chemical exposure stores an offset that calibration cannot clear. That is why a swap restores the reading at once and why power cycles change nothing.

A second step at restart, from 40% down toward 10% and climbing back over hours, fits a film that re-equilibrates to its own wetting history before it tracks the air again. The cost of trusting such a trend is easy to picture: chasing a displayed 40% against a true 70% drives a humidifier until walls and stock stay damp.

What do the symptoms tell you before you swap parts?

Four trend patterns sort the causes: a fast step that never recovers points to a contamination event, a slow creep points to aging, a reading that returns after time in dry air points to residue that can desorb, and a low reading after cold nights points to condensation.

• A step of 10 points or more within minutes, flat for weeks. A contamination event: condensation dried on the film, or a chemical release nearby.
• The correct reading returns after weeks out of service. Reversible residue; the vendor bake compresses that recovery into 10 h.
• A creep of a few points over a year. Normal polymer aging; correct the offset in software and re-test annually.
• Low after cold nights, back when warm. Condensation on the element or the housing wall; fix the housing, not the sensor.
• Jumps of 20 points or more with no pattern. Look at wiring and connectors first: a heat-shrunk joint that opens with temperature mimics drift.

What contaminated the element, in order of likelihood?

Three causes cover nearly every low-reading failure: first condensation that dried and left dissolved residues on the film, then plasticizers and solvents outgassing from nearby plastics and adhesives, and last salt or cleaner deposited onto the element or its membrane.

1. Condensation with something dissolved in it. Clean water dries off a film without a trace; the damage comes from what the droplet carried. In a coastal animal room holding 60% RH while temperatures cycle 23 to 27 °C, the dew point sits near 18.6 °C. Any wall that dips below that overnight grows a droplet loaded with salts, ammonia or cleaning agents, and the residue stays on the film when it dries. A fully sealed housing makes this likelier, because each temperature swing pumps moist air through gasket gaps; the trade-offs of humidity sensors inside waterproof housings decide which side of that you want.

2. Outgassing from nearby plastics. Adhesives, conformal coatings, fresh heat-shrink and cable-tie compounds release volatile organics that a polymer film absorbs; an application note on chemical contamination from a sensor-chip maker lists adhesives and coatings among the usual culprits. A step within weeks of a rework job points here. Keep bare elements away from fresh joints, and let new heat-shrink finish off-gassing before the housing closes.

3. A direct film on the element or its membrane. Salt spray, washdown mist or fine dust packed onto the sensor's protective membrane blocks vapor and holds a damp micro-climate against the film. Compare the element with a fresh part under a loupe: a clean sensor shows a bare, even surface, a fouled one shows white or brown patches.

Lifetime follows the humidity the film actually sees. Below about 90% RH with condensation excluded, clean polymer parts hold their few-percent accuracy for years; with condensation or contaminants in the picture, months are realistic.

How do you check whether condensation reached the element?

Work out the dew point before you blame the wiring: at 27 °C and 60% RH it sits at 18.6 °C, so every interior surface below that collects condensate, and one night of wall temperatures tells you whether the wetting event happened.

• Swap test. Run the suspect part beside a trusted meter for 24 h. A suspect alone 30 points low has a fouled or altered film; both meters low together means the room really dried. A bad joint steps with handling, not with weather, so this test clears wiring too.
• Dew point check. The Magnus approximation gives the saturation temperature of your air. First ln(0.60) = −0.511 and 17.62 × 27 ÷ (243.12 + 27) = 1.761, so γ = −0.511 + 1.761 = 1.250. The dew point is 243.12 × 1.250 ÷ (17.62 − 1.250) = 303.9 ÷ 16.37 = 18.6 °C [65 °F]. A wall at 16 °C sits 2.6 °C below that and condenses on every cold night; a wall at 20 °C carries a 1.4 °C margin and stays dry. Tape a cheap thermocouple to the coolest interior wall for one night cycle and you have your answer. Condensation on cold surfaces in a humid room follows the same mechanism one scale up.
• Installation history. A step within weeks of rework, such as new heat-shrink or adhesive near the element, points to outgassing rather than water.
• Loupe inspection. White or brown film on the element, next to a clean new part, confirms a deposited residue.

Can you bake a drifted sensor back to spec?

The vendor handling instructions give the recovery bake: 100 to 105 °C [212 to 221 °F] at under 5% RH for 10 h, followed by re-hydration at room conditions, then a comparison against a trusted meter before the part goes back on trend.

The bake drives absorbed water and volatile contaminants out of the film. It is the same desorption that rescues a part left on a shelf for weeks, compressed into hours. Re-hydration re-balances the film, so the first readings after the bake are not the judge. On a populated board, run the bake only when every part on it is rated for 105 °C; otherwise move the sensor to a bare test board for the 10 h.

Then re-test beside the reference meter for 24 h and decide from the table in the last section. Skip solvent rinses entirely: the rinse is itself a contamination event for a bare film. The same dry-it-or-replace logic after any wetting event is walked through for displays in condensation inside an outdoor HMI.

How do you mount the sensor so it stays accurate?

Give the film room air without room water: a pressure-equalization membrane vent on the housing, the element clear of cooling walls and fresh adhesives, and a local RH that stays under about 80% through the night cycle.

• Vent the housing. A membrane vent stops the pressure pumping that drags moist air through gasket gaps, passes vapor so the element tracks the room, and keeps liquid water out. Pick a vent rated to match the housing, because the weakest part sets the assembly's rating; our housings are rated by the manufacturer, not UL listed. Choosing a pressure equalization membrane vent covers the sizing.
• Warm beats cold. Air at the same moisture reads lower RH when warmer. Room air with an 18.6 °C dew point reads 60% RH at 27 °C and about 45% at 32 °C: saturation vapor pressure climbs from 3.57 to 4.76 kPa, and 60 × 3.57 ÷ 4.76 = 45. The element's own heat is no factor, since it draws about 3.6 µA at 1 Hz; the enclosure relative-humidity calculation shows what the film sees when the housing runs warmer than the room.
• Add heat where the room stays wet. Where the room holds 85% RH or more for weeks, a sensor heater that keeps the element a few degrees above the dew point is the established fix, and the humidity sensor heater strategy gives the control logic.
• Point the entries down. Glands and cable entries face down with a solid top, so a drip on the outside cannot reach the film; the wall-mount orientation rules apply to sensor pipes too.

When do you replace the sensor instead of trusting it?

Replace the part when a bake-and-re-hydrate cycle still leaves it more than about 5 points low against a trusted meter, when the drift repeats within months of a clean installation, or when the film has been wet with anything beyond clean water.

Rank the causes the same way every time: a condensation event that dried and left residue first, outgassing from a recent rework second, a direct spray film last. Then fix in that order — bake or replace the part, fit the vented and warm mount, and only after two failures in eight months in one housing re-spec the mount itself: vented, heated or relocated, not a third sensor.

Catalog dimensions to compare

Compare these catalog dimensions with your component layout. Open each enclosure for its rating, material and usable space, then check those specifications against your application.

Questions