A hygrometer reading is relative humidity, which is a ratio and not an amount

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

The reading is a ratio, not an amount of water

Air has a ceiling on how much water vapour it can hold, and that ceiling climbs steeply with temperature. In the ordinary room range it roughly doubles for every eleven degrees Celsius. A hygrometer compares the vapour present against that ceiling and reports the result as a percentage, so the number has a temperature baked into it that is not displayed as part of the humidity figure.

The consequences are not intuitive. Warm the air and the reading falls. Cool it and the reading rises until it hits a hundred and water starts condensing out. Move the sensor nearer a radiator and it reads lower; put it on a cold outside wall and it reads higher. In every one of those cases the amount of water in the room is identical. Only the denominator moved.

This is the whole content of the pillar page on what relative humidity is, and it is worth internalising before trusting any percentage, because almost every surprising humidity reading turns out to be the denominator rather than the numerator.

Same water vapour, three different percentages

Holding the water content fixed and varying only the temperature makes the point arithmetically. Take air carrying 9.4 grams of water vapour per cubic metre, which is a perfectly ordinary indoor quantity, and read it at different temperatures.

Five very different readings describing one unchanged parcel of air. If you want a figure that does not do this, use dew point, which stays at 10 degrees C across every row of the table, or a mixing ratio in grams per kilogram of dry air, which is also conserved when you heat or cool the air. Relative humidity is the right quantity for comfort, mould and material equilibrium, because all three respond to the ratio rather than the absolute amount. It is the wrong quantity for asking whether ventilating actually removed water.

How a polymer film becomes a percentage

Inside almost every modern hygrometer is a capacitive element: a thin hygroscopic polymer film sandwiched between two electrodes on a silicon die. Water molecules diffuse into the polymer, raise its dielectric constant and therefore the capacitance of the sandwich. The chip measures that capacitance, measures its own temperature, and uses a factory determined characterisation to convert the pair into a percentage.

Three things follow from that construction, and they explain most of the complaints people have about hygrometers.

• Every humidity reading is two measurements. An error in the temperature channel propagates straight into the humidity figure. This is why a sensor warmed by its own electronics reads low on humidity even when its temperature number looks plausible.
• Diffusion takes time. The film has to exchange water with the air through whatever housing and dust filter sit in front of it. Response is measured in tens of seconds to minutes in still air, far slower than the display refresh.
• The film has a history. It does not return from damp air as quickly as it went in, and prolonged time near saturation shifts the characterisation itself.

The sampling interval is independent of all three and often confused with them. The Govee H5074 Bluetooth thermometer hygrometer states a reading every two seconds and a tolerance of plus or minus 3 percent relative humidity in the same bullet, which is a neat illustration: one of those numbers is about how often you see a figure and the other is about how much that figure can be wrong, and they are unrelated.

Hysteresis and the slow walk back from saturation

Approach 60 percent from a dry room and a capacitive sensor settles within a couple of minutes. Approach the same 60 percent after two days in a bathroom at 90 percent and it can sit a point or two high for hours. That path dependence is hysteresis, and it is a property of the polymer rather than a defect.

Extended exposure well above about 80 percent relative humidity does something more lasting, usually called saturation drift: the film’s response curve itself shifts, so the sensor reads consistently high even after it has dried. Some sensors recover with a controlled reconditioning at elevated temperature and low humidity. Many consumer units simply carry the offset permanently, which is a strong argument against leaving a cheap unit in a greenhouse or a bathroom and then trusting it elsewhere.

The practical response is to establish an offset and recheck it periodically rather than assuming the factory characterisation survives. The salt slurry method in calibrating a hygrometer is the standard way to do it, and a unit with a replaceable sensing element gives you a second option when the drift is too large to correct.

Where a published accuracy stops applying

A tolerance usually applies over a narrower span than the measuring range, and the two numbers sit in different bullets. The HOBO temperature and humidity USB data logger is unusually explicit about it: its published specification states a relative humidity range of 15 to 95 percent and, separately, an accuracy of plus or minus 3.5 percent relative humidity from 25 to 85 percent. Between 15 and 25, and again between 85 and 95, the instrument still displays and records a number with no stated tolerance attached to it.

That is not a criticism of the logger, which is being more forthcoming than most. It is the cleanest available illustration of what a humidity reading is: a converted figure whose trustworthiness is a separate, bounded claim. The same listing also states a user replaceable humidity sensor, which is the manufacturer acknowledging that the element drifts.

Compare that with what other listings publish. The Inkbird ITH-10 thermometer and hygrometer states that it records maximum and minimum temperature and humidity and switches between Fahrenheit and Celsius; a humidity tolerance is not stated in the listing bullets. The Freshliance Bluetooth temperature and humidity data logger states app based analysis with maximum, minimum, average and mean kinetic temperature values computed from the record, and no humidity tolerance figure appears either. Mean kinetic temperature is worth noting as a specifically thermal construct; there is no humidity equivalent of it in that bullet, so the humidity channel is being logged rather than summarised.

Turning a percentage into something actionable

Once you treat the reading as a ratio, a few habits follow naturally and most of the confusion goes away.

• Always read the temperature next to it. A humidity figure without its temperature is half a measurement, and the two together let you derive dew point, which is the figure that predicts condensation.
• Compare readings at the same temperature or not at all. Two rooms at 55 percent and 20 degrees C hold the same water. Two rooms at 55 percent and ten degrees apart do not.
• Keep powered sensors out of their own heat. Away from a wall socket, off the top of an appliance, out of direct sun. Self heating biases the ratio downward in a way no calibration corrects.
• Give it time after moving it. Diffusion through the housing means the first several minutes describe where the sensor came from.
• Judge the trend, not the digit. A fixed offset cancels when you compare this week against last week, which makes a logged series far more useful than a glance at a display.
• Check what the tolerance covers. If your target band sits outside the span where the tolerance applies, the instrument is not specified for the job you bought it for.

If you are choosing an instrument with those habits in mind, the selection pages for humidity meters and combined temperature and humidity monitors separate the units that publish a tolerance and a span from the ones that publish only a refresh rate.

Frequently Asked Questions