One of my hobbies is doing crazy projects — it seems the more useful and practical the project, the less likely I am to be interested in doing it. One project I’m working on is building a humidity sensor from scratch — based on the chilled mirror principle.
Chilled mirror dew point sensors
These sensors work by cooling down a mirror until it starts to form condensation (i.e., “fog up”). If the temperature of the mirror is below the dew point of the air, condensation will form. The dew point is a function of the absolute humidity of the air — the higher the humidity, the higher the dew point. So, a control system lowers the temperature of the mirror if it’s free of condensation, and raises the temperature if condensation starts to form. The system ends up settling in a state right on the borderline, with just a tiny amount of condensation that can only be maintained by keeping the mirror ‘s temperature locked to the dew point of the air. A sensor then measures the temperature of the mirror, which can be used to calculate the humidity.
The system keeps track of the amount of condensation by bouncing light off the mirror. If there’s condensation, some of the light will be scattered away, and a light sensor can measure the scattered light (or the corresponding dimming of the reflected beam). When a lot of light is scattered, there’s too much condensation, and the mirror is heated up. Too little light being scattered means the mirror needs to be cooled down. The heating and cooling are often done using a thermoelectric cooler module (also known as a Peltier cooler).
This technique allow measurements typically accurate to ±0.5% relative humidity, although this level of accuracy requires frequent cleaning (according to an uncited Wikipedia article — yes, just about everything on this post is straight off Wikipedia). The main advantage is that chilled mirror sensors measure the dew point of the air in a fundamental and direct way, so they’re not as susceptible to aging, drifting and contamination. as other techniques.
Capacitors or resistors can be made using special dielectrics or resistive elements whose values vary according to humidity. These are comparatively cheap and easy to mass-produce, but are susceptible to aging effects and contamination. They can have an accuracy of ±2% relative humidity, but only after calibration — which raises the question, how do you calibrate it? (answer: often, with chilled mirror sensor.)
Other non-electronic methods involve a pair of thermometers, one with a wet cloth on the bulb. By passing air over the wet cloth, the moisture evaporates into the air, cooling it down. The humidity can be inferred from the two temperatures. Yet another method uses an animal hair, which changes length according to humidity. This can be used to move a dial.