The traditional hair, of horse or human, is still used in high quality industrial RH sensors. The movement of the hair is turned into an electrical signal by a strain gauge (described elsewhere in this rambling book). Similar sensors are made of a strip of cellulose butyrate, a water absorbent polymer which likewise stretches and shrinks according to RH.
The sensor type that dominates the transportable data logger and hand-held instrument market is, however, the capacitive sensor.

The capacitive RH sensor consists of a thin layer of water absorbent, polymeric or inorganic material, that is coated onto a conductive base. The layer is then covered with a porous conductive layer. As the relative humidity increases the water content of the polymer increases. Water has a high dielectric constant. This means that the combination of the two electrodes with the water between can store more electric charge. This electrical capacity is measured by applying a rapidly reversing (AC) voltage across the plates and measuring the current that passes. Note that the polymer or inorganic material (often aluminium oxide) plays only an indirect part in the measurement: it is the abundance of water molecules that is measured.
The device is typically 7 x 4 x 0.5 mm thick.

The principle is simple enough, but there is a long history of development of practical sensors that are resistant to air pollutants, or immersion in liquid water. I do not recommend making one from first principles, but read on if you are a do-it-yourself enthusiast.

Capacitive sensors have a serious limitation: the change in capacitance is small compared with the capacitance of even a few metres of cable. This means that the electronic processing has to be completed close to the sensor. If one data logger is connected to several RH sensors, each will need its own power supply (extra wires in the cable) and relatively bulky electronics. This is an irritating extra expense, because some data loggers have all the necessary processing power built into them.

The resistive RH sensor is a thin wafer of water absorbent polymer, printed with two interlocking combs of conducting metal or carbon.

The device illustrated is from General Eastern. It is about 10 mm long. It can be bought alone, without the processing electronics

The quantity that is measured is now the simple electrical resistance through, or across the surface of, the polymer, which changes with water content. This sensor, curiously enough, also needs an alternating excitation voltage, not for the measurement but to avoid destroying it by causing one-way electrolytic ion movement in the polymer. The Campbell data logger, mentioned earlier, provides a short voltage pulse, then pauses before sending a negative pulse of the same magnitude. The direct current in each episode is measured and used to calculate the RH.

I mention this intricate detail because many data loggers cannot provide this excitation.

The resistive sensor has dropped out of fashion, though it is still used in the "Trend Reader" data logger. The principle of measuring the resistance of a water absorbent polymer is, however, well suited to do-it-yourself manufacture and such devices have been used in high quality research, mostly in building studies where rapid response is not important.

The first of these slow and simple devices, as far as I know, was the matchstick (de-waxed) coated with silver paint on two sides, known as the Duff sensor.

The two wires shown in the diagram are glued into grooves in the wood. The exposed half of the wire is then filed flat. Silver, or carbon, paint is then applied over the two surfaces with embedded wires. In this "refined matchstick" model the end grain of the wood is exposed on the long, uncoated sides to give a faster response.

The same principle of measuring the resistance of a relatively thin layer of wood has been applied even more simply by banging two stainless steel nails about 3 mm apart into a cork shaped block of wood. The problem here is keeping the nails parallel. A pre-bored hole helps. These devices take several hours to reach equilibrium.

The down side to these cheap and simple sensors is that they require correspondingly complicated data loggers that are capable of providing the pulsed excitation.

They have their uses in places such as museum stores, which are often so stable that there is no need for rapidly reacting sensors. In archives the objects are often boxed, so that short cycles in RH will be "buffered out". The running average over an hour or two, given by a small piece of wood, is perfectly adequate.

Often the real need in stores is for remote detection of catastrophe. The same resistance measurement can double as a flood detector. The bared lead wires to the RH sensor are draped closely parallel just above the floor for a metre or two. Flood water is usually a good electrical conductor, for unpleasant reasons. A more elegant water detector is described later in connection with the weather station.

RH sensors are delicate. They are easily contaminated by water soluble salts, which are nearly universal in walls. The sensor can be protected by sealing it in an envelope of Goretex or of thin (less than 0.5 mm) silicone membrane. These materials stop ions while allowing water molecules through to influence the sensor. The envelope should not touch the sensor surface. The leads are difficult to seal. If there are two separate wires they should be spread away from each other to lessen the chance of a hole that the sealant (silicone) cannot reach. These semi-permeable membranes will lengthen the response time of the thin film sensors from about a minute to about twenty minutes.

Calibrating RH sensors

RH sensors are notoriously unreliable. They MUST be calibrated on installation and about every six months thereafter. Calibration involves checking the accuracy at at least two RH values. The usual method is to use saturated salt solutions.

Light and uv sensors


Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.