Skip to content
AkerminMeasure. Monitor. Improve. Compare Products
Learning Center

Can a hygrometer be calibrated, or only checked and corrected?

Calibration means adjusting an instrument against a reference. Verification means finding out how far out it is and living with the answer. Almost every household humidity sensor permits the second and not the first, and the salt that makes it possible costs nothing.

By the Akermin Editorial Team · 6 min read

As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.

Calibration in the strict sense means comparing an instrument against a traceable reference and then adjusting it so that it agrees. Almost no consumer hygrometer permits the adjusting half, because there is nothing to turn and nothing in the firmware to change. What every hygrometer permits is verification: put it in air held at a known humidity, record what it says, and from then on carry the difference in your head or on a label. The known air is the easy part. A saturated solution of a common salt in a sealed jar holds the space above it at a fixed and published humidity, which gives you a reference point for the price of a spoonful of salt.

Calibration, verification and an offset field are three different things

These words are used interchangeably in product copy and they mean distinct operations. Calibration compares and adjusts, and it leaves the instrument reading closer to truth than it did before. Verification compares and records, and it leaves the instrument exactly as it was while leaving you better informed. An offset field in an application compares nothing at all; it adds or subtracts a figure you typed, after the measurement, and the quality of the result depends entirely on the quality of the figure you supplied.

The distinction has practical bite. A verified sensor reading 71 where the reference says 75.5 is a usable instrument, because you now know to add four and a half points in that region. An unverified sensor with a confident app offset applied to it is worse than an untouched one, because the correction is invented. And a laboratory calibration certificate, which is the only thing that properly deserves the word, would normally cost more than the class of instrument it was issued for, which is why the household version of this job stops at verification. The step by step handling for the check is in how to calibrate a hygrometer.

A saturated salt holds the air above it at a known humidity

The physics is worth stating because it is what makes the method trustworthy. In a sealed container holding a saturated salt solution with undissolved salt still present, the solution cannot become more or less concentrated: take water out of the air and some salt precipitates, put water in and some salt dissolves. Concentration is pinned, so the vapour pressure above the solution is pinned too, and the relative humidity of the enclosed air settles at a value that depends only on the salt and the temperature. Those values have been measured carefully and published for decades.

Saturated salt Nominal RH at 20 degrees C Nominal RH at 25 degrees C
Lithium chloride 11.3 percent 11.3 percent
Magnesium chloride 33.1 percent 32.8 percent
Magnesium nitrate 54.4 percent 52.9 percent
Sodium chloride 75.5 percent 75.3 percent
Potassium chloride 85.1 percent 84.3 percent
Potassium sulfate 97.6 percent 97.3 percent

Read down the two columns and notice how differently the salts behave with temperature. Sodium chloride moves about two tenths of a percentage point across five kelvin, which is negligible against any consumer tolerance. Magnesium nitrate moves a point and a half over the same interval, so a check with it is only as good as your control of temperature. That is the main reason ordinary table salt is the recommended first reference: it is cheap, it is stable, and at roughly 75 percent it sits in the band where mould and condensation questions are actually decided.

The one point check, and what one point cannot tell you

Mix salt with a little water into a slurry, not a solution, so that undissolved crystals remain visible. Put it in a sealed container with the sensor above it and not touching it. Leave it somewhere with a steady temperature for eight to twelve hours, then read the display without opening the container. Subtract.

What you have learned is one number at one humidity at one temperature. That is genuinely useful and it is also the limit of the exercise. You do not know whether the error you found is an offset that applies everywhere, a scale error that grows with the reading, or a local distortion. You do not know the time constant of the sensor, which matters if you are watching a room that changes quickly. And you do not know whether the figure would repeat, because a single check has no repeatability information in it at all. Two salts give you far more than twice as much: with 33 percent and 75 percent you can see whether the two errors have the same sign and the same size, which is the difference between an offset you can safely apply and a slope you cannot.

Record the result where it will survive. The Govee H5074 states twenty days of onboard storage, a curve graph and CSV export of the last two years to a phone, which makes the check auditable: you can see the whole equilibration curve rather than one final reading, and a curve that was still climbing when you read it tells you to wait longer next time. A stated accuracy of plus or minus 3 percent RH is the band the verification is testing, and the JEDEW hygrometer review works through what such a band permits before any correction is applied.

Why a sensor that reads low at ninety may read high at thirty

Capacitive polymer sensing is not linear and its errors are not uniform across the range. Three mechanisms matter. Hysteresis means the film gives water back more slowly than it takes it up, so a sensor arriving from a humid room reads differently from one arriving from a dry room, and the gap can be several percentage points for hours. Contamination means dust, cooking aerosols, solvents and smoke deposit on the film and change how much water it holds at a given vapour pressure, which usually shifts the top of the range more than the bottom. Ageing means the polymer itself changes, slowly and in a direction that depends on what it has been exposed to.

Because those act differently at different points, the shape of the error curve changes rather than the whole curve sliding. That is why extrapolating a correction found at 75 percent down to 30 percent is unsafe, and why manufacturers quote a single tolerance band across the range instead of a correction table. Several of these listings illustrate what the published figure covers and what it leaves out. The GoveeLife gateway sensor pack states plus or minus 3 percent RH with readings every two seconds and threshold alerts by app and email, with whether an offset can be entered Not stated in the listing bullets. The Temtop S1plus states that Bluetooth connectivity supports firmware updates over the air, which is sometimes misread as a calibration route and is not one; a firmware update changes software, not the sensing film. The ALLOSUN thermometer hygrometer review covers the harder case, an instrument that publishes no humidity tolerance for a verification to be measured against.

What to write on the back of the sensor

A verification that lives only in your memory has already expired. Put a small label on each unit carrying the reference humidity you used, what the unit read, the temperature you held, and the interval you allowed. Repeat the check annually, and immediately after any event that could have contaminated the film, such as a condensation soaking, a paint job in the same room or a move from a cellar to a kitchen. Reconditioning, meaning a long spell in dry air before re checking, will recover some sensors that have been sitting near saturation and will do nothing for one that has genuinely failed. The decision rule is simple enough: a unit whose error is small and repeatable is a working instrument with a known correction, and a unit whose error is large or wanders between checks is not an instrument at all.

Frequently Asked Questions

Which salt should I use for a single check?

Sodium chloride, ordinary table salt, because it holds the air above a saturated slurry at about 75 percent relative humidity and is the least temperature sensitive of the common references. It is also the band where damp problems live.

How long does a salt check take?

Plan on eight to twelve hours in a sealed container at a stable temperature, and ignore anything you read in the first two. The sensor and the air both have to reach equilibrium, and a polymer film that has been in a dry room takes hours to take water back up.

Can I just subtract the error I found?

Only near the humidity you checked at. Capacitive sensor error is rarely a constant offset, so a correction established at 75 percent may be wrong in both size and direction at 30 percent. Two reference points tell you whether the error is an offset or a slope.

Do smart sensors have a calibration setting?

Some apps expose an offset field, but none of the four listings behind this page states one. Treat an offset field as a note to yourself stored in software rather than as a calibration, because it does not change what the sensor does.

What if the reading is far out at the reference point?

A sensor that reads ten or more percentage points away from a well made reference has either been contaminated or has failed. Try a reconditioning cycle in dry air, repeat the check, and if it has not moved, retire the unit rather than correcting it.