Harrison’s bimetallic strip and the problems with temperature control

What does a thermostat do? The question underscores why thermodynamics is such a pig to learn. Not only is the subject complex and mathematical, but its concepts are often quite at odds with our lived, sensory experience.

So maybe I should rephrase the question. You are sitting at home and you feel a little cold. What do you do? Some might get up and don another layer. But imagine that you have a radiator nearby, and it is equipped with one of those valves – a thermostatic valve – that are helpfully labelled with numbers from 0 to 5. Do you adjust it? Why would you do that? And if you do, what setting do you choose?

From my experience, many twist the valve to the highest setting. Even in rooms with digital thermostats mounted on a wall, it is very common for people who feel cold to increase the temperature setting – I have found settings of 27˚C, a temperature that in summer would cause many to complain that it is too hot to work. If questioned about their choice, they will reply that they ‘feel cold’ and want ‘more heat’ to come from the radiator. In an instant we are up against three cornerstone concepts of thermodynamics: temperature, heat and power.

Their conflation is neatly summarised by the question ‘Why does a person make a poor thermometer?’ that Henry Bent put in his thought-provoking textbook The Second Law. This stems from overlapping perceptual illusions. Our perception of temperature is a complex neural construct that is a product of our life history. Whether we feel warm or cold will depend on our age, our health, how we grew up and where we have just been. Come in from a cold street and you might think a room hot; others who have been there for a couple of hours will say the opposite. No pointing at the reading of a thermometer will make a jot of difference to their complaint.

Our societal and cultural norms add further layers to this issue. In her book Invisible Women Caroline Criado-Perez draws attention to the thermostat wars that are fought in offices between men and women, driven in part by physiological differences, but also by the fact that women are often expected to wear flimsier, and fewer, garments at work than men. And at the same time, almost all the controls we twist – think the volume control on a stereo or the flame on a gas cooker – are, literally, knobs of power that directly control the rate of energy transfer.

Switching on

So what is our thermostat doing? The answer is oddly prosaic: it’s nothing more than a switch that responds to temperature – independently of our perception. It will allow hot water or an electrical current to flow, or a fan to blow, provided the temperature is above or below a certain limit. How much heat the convector in your home heating delivers has nothing to do with the thermostat. Instead it depends on the temperature of the fluid coming from the heat source (whether electrical element, gas boiler or heat pump). Turning the knob to 11, Spinal Tap-style, simply means that your device will be permanently on and uncontrolled, but not hotter. It is ironic then that advertisements of thermostatic radiator valves always show them set to bright red ‘MAX’, underscoring the misunderstanding of what thermostats are for, and leading to enormous energy waste.

The need to control temperature goes back to time immemorial. This might have involved tossing another log on the fire, or coal into a furnace. Alchemists needed gentle control for ‘cohobation’ (stewing) of certain mixtures and stronger heat for distillation, traditions that have strong parallels in modern cooking where cooks will add water or wine to moderate temperatures in a pan. Our instinct drives us to seek a violent rolling boil to our pots and flasks, when as scientists we know that a liquid is unlikely to warm by more than a few degrees above its boiling temperature – neither your pasta nor your prep will be done any sooner.

But smelting, firing pots or making glass required air temperatures inside ovens to be steady and reproducible, long before temperature was something one might measure. It was the link between expansion and temperature that made measurement possible and soon after, control too.

The earliest temperature regulator was invented by the Dutch polymath and inventor Cornelis Drebbel (1572–1633), who also constructed a wooden submarine that is said to have impressed James I by travelling underwater from Greenwich to Westminster. Among his myriad inventions was what Drebbel called a Perpetuum Mobile, a peculiar device consisting of a hollow globe connected to a circular glass tube around its outside. Changes in temperature expanded and contracted the air in the sphere, pushing water in the circular tube to move back and forth ‘like the tide’ or even rotate.

Drebbel applied this principle more practically to a self-regulating oven where the expansion and contraction of an alcohol and mercury piston moved a baffle that checked the air entering the furnace. Drebbel’s son-in-law made an incubator to hatch eggs at any time of year, ‘just as if they had been hatched by ducks or hens’, wrote one observer.

Links to longitude

But the need for temperature compensation in clocks led to a simpler device, again based on thermal expansion, and one that is everywhere in our lives. In the 1700s both the French and English governments established prizes to solve one of the key problems of navigation that dogged their countries’ commercial fleets: longitude. Without a reliable clock, it was impossible to establish the east-west location of a ship on the ocean and such errors resulted in enormous delays and financial losses for firms trading across the globe.

As described in Dava Sobel’s wonderful book Longitude, self-taught provincial clockmaker John Harrison solved the problem with a series of clocks of revolutionary design. At their heart were inventions designed to counteract changes in temperature. Harrison’s first device was a pendulum in which a perpendicular grid of brass and steel kept the weight swinging at a constant distance regardless of temperature. This was replaced in his later clocks by two slender strips of brass and steel riveted together that curled gently one way or another to compensate for changes in temperature.

Harrison’s empirical approach was updated by the meticulous mathematical approach of French astronomer and mathematician Yvon Villarceau (1813–1883). He included a detailed method for making fused steel/brass strips. With this new predictability, strips began to find ever more uses. Curved into spirals like the tube of a Bourdon gauge they could unfurl with temperature, driving a needle on a display. Cheap, robust and mercury-free thermometers could measure temperatures from steam engines to refrigerators. But a similar spiral design could be used to open and close valves both on engines and on heating systems. You probably have them in the mixer taps of your sinks and showers.

As digital controls expand into our lives, they become ever more sophisticated – nowadays it is not just furnaces that use PID algorithms, but home heating systems too. But are they really better than Harrison’s bimetallic strips? The irony is that for all the mechanical or digital sophistication of our controls, in the end how we use them is what really matters, and that is down to our own sensual fallibility. Therein lies a large part of the challenge of addressing our energy security.

Acknowledgments

I am grateful to Michael de Podesta for myriad discussions and for applying a fine razor to this essay.