The next topic is the sensors that are connected to those data loggers that come with vacant screw terminals instead of built in sensors. The huge advantage of the programmable electronic loggers is the variety of sensors that can be attached; the disadvantage is learning about the sensors and the programming.

The thermohygrograph and most of the ready-to-go electronic loggers are limited to measuring the climate in the main body of air in a room, or maybe a large showcase with well oiled door hinges. Their bulk prevents measurement of conditions close to an outer wall, for example, or at the surface of a painting under intensive care on the low pressure table. This is a serious limitation to the diagnostic usefulness of the device. The boundary layer of air next to a wall, that is the layer in which the entire transition from the room climate to the wall surface climate occurs, is usually less than a millimetre thick. This layer is often important. The prevalence of mould in the corners of rooms, for example, is usually due to the low temperature at the wall surface. One talks loosely of "damp corners" but usually the moisture content of the air is the same in the corner. It is the lower temperature that causes a higher relative humidity. Many people have sought, or theorised over, unusual sources of water in such places when the source is simply the air in the room.

This sort of investigation can only be done with an electronic sensor that is small enough to measure the surface temperature. It is this kind of investigation that the electronic data logger is peculiarly well suited to.

Types of temperature sensor

There are four commonly used temperature sensors.

The semiconductor sensor measures the leakage current through a semiconductor, which is proportional to the absolute temperature. This is not much used in general data gathering. It is not very small and has no particular advantages.

The platinum resistance sensor changes resistance according to the temperature, calibrated by a very well defined equation. It is often used as an accurate reference but, dare I say, this accuracy is seldom needed in our trade.

The thermistor is a semiconductor which also changes resistance with temperature, but much more than platinum (and in a different direction). It is now the most commonly used electronic temperature measuring device, after a hesitant start when its reliability was not so good. The thermistor is quite small, typically a 2-3 mm round blob on the end of two wires. Since this is thicker than the boundary layer it is not quite suitable for measuring the temperature of window glass or damp walls.

The thermocouple uses the phenomenon that a voltage is spontaneously generated between the ends of an electrical conductor that passes through a temperature gradient. If two different conductors pass through the same temperature gradient, then the voltage developed will be different in each. The voltage difference is forced to zero at the remote end by welding the wires together there, so the voltage that appears between the wires at your end can be translated into temperature difference between the two ends.

Because the thermocouple only detects a temperature difference, a thermistor (usually) at the voltage measuring instrument defines the temperature there, from which the temperature at the remote end can be calculated.

This seems an awfully complicated way of making a simple measurement. There are two reasons why thermocouples are widely used: the irrelevant one, for us, I hope, is that they can operate in high temperatures that would destroy the other sensors mentioned. The more interesting reason for us is that thermocouples can be made extremely small: as small as one can weld two wires together, which is about a tenth of a millimetre for average skilful fingers. They are very cheap, particularly if you solder the wires yourself, which is good enough for room temperature investigations.

A less obvious technical advantage is that they are voltage generators with a very low internal resistance, which makes them resistant to electrical interference even over long cable runs.

A thermocouple hardened for installation outdoors. The colour coding of thermocouples is bright and cheerful, but not internationally standardised. This is American type K. The wires are bared, twisted and then hard soldered. A heat shrinkable tube (blue) is slipped over the round teflon outer sheath. The space around the junction is filled with magnesium oxide powder (white), to absorb water diffusing into the assembly. A stainless steel plug (cut from a welding rod or bicycle wheel spoke) is pushed in and then the tubing is heated to shrink it snugly around the assembly. Thermocouples can also be bought in stainless steel sheaths.

Types of thermocouple

The wire pair recommended for general use at natural temperatures (-70 to 100 degrees) is copper-constantan (which is a copper nickel alloy). This is known in the trade as type T. There is one important disadvantage to this type: the thermal conductivity of the copper wire is so good that there is a risk of heat transfer to the measuring tip. This is particularly annoying in measurement of the surface temperature of relatively poorly conducting wall surfaces. The temperature measured is somewhere between the true surface temperature and the temperature in the room, which is "leaking" along the copper wire to the tip. If this error matters it is better to use type K: chromel-alumel. This is generally regarded as less reliable than type T, because small variations can be caused in manufacture or by straining the wires during use or installation. I have never noticed errors of this nature that come close to the errors caused by heat transfer along the wires, so I recommend type K for investigations of small scale climate variation.

Next part: Relative humidity sensors for electronic data loggers


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