In-Depth: Consider The Mainspring
One of the most essential, and most underappreciated, components in watchmaking.
Despite the fact that I have been seriously interested in watches and watchmaking for probably thirty years or so, I have given very little thought, in all that time, to the mainspring. This seems a serious deficiency as of course, without mainsprings, there are no watches and there is no watchmaking.
Its essential nature notwithstanding , the mainspring is probably the least sexy part of a watch to just about anyone interested in watches – a positive horological anaphrodisiac, as it were. Different folks find different things interesting, of course – some people love minute differences in dial lettering; some people find escapements endlessly fascinating; some people love the intellectual challenge of understanding how various complications work, and on and on. However, though the mainspring is probably last on any watch enthusiast’s list of things of interest about watches, without them virtually all the watches we love could not exist (G-Shock collectors are off the hook).
I became curious about mainsprings because, having recently found myself with a little extra time to consider things, I had been considering steel. This was sparked by a series of long conversations I’d been having with my older son, about why there was no industrial revolution in ancient Greece or Rome, or in China or anywhere else. As Jared Diamond famously pointed out in Guns, Germs, And Steel, it actually takes quite a lot of factors all coming together in just the right way, for an industrial revolution to occur. Everything from advances in basic science, to the evolution of mathematics, some degree of social stability, and various other discoveries all have to take place, and in conjunction with each other. (The ancient Greeks, for instance, likely had rudimentary steam engines and certainly had an understanding of gearing and complex mathematics, but the lack of more tractable mathematics tools as well as the inability to produce metal alloys of reliable quality in bulk, are just some of the reasons why Plato never rode a steam-powered Athens-To-Sparta Limited, running on iron rails). The ability to produce steel in quantity is one such essential element – to all forms of industry, and of course, watchmaking.
Steel itself has been known since antiquity. The Romans used it for swords; high-quality steels were made in China as early as 400 BC; and later, parts of the world as varied as India and Japan became famous for steelmaking by the Middle Ages. The salient characteristic of these methods of steelmaking, however, was that they relied on carefully transmitted rules of thumb, rather than an understanding of basic chemical metallurgy. Steelmaking for most of human history was a small-batch, artisanal process rather than an industrial one.
The production of large quantities of iron of predictable properties, and later steel, was indispensable to the industrial revolution as it took place in Europe, and these technical advances took place quite late in history, relatively speaking. The transition took place over the course of the 19th century, when we went from being able to produce some cast and wrought iron, as well as relatively small amounts of steel, to being able to produce various grades of iron, and high-grade steels, in really industrial quantities (stainless steels were the latest to the party; but by 1908, a sailing yacht of stainless steel weighing over 300 tons had been built).
The reason all this is relevant to watchmaking is that spiral steel springs were indispensable to the manufacture of portable timekeeping devices. The earliest known clocks in Europe were not powered by mainsprings. Instead, they were driven by weights: you put something suitably heavy on the end of a rope wound around a pulley, and as gravity pulls the weight downwards, the linear movement of the weight is converted to rotational torque which can be used to drive a gear train. The idea is so simple that it seems impossible that it didn’t occur to someone in the ancient world – after all, the folks who could produce things like the Antikythera Mechanism were clearly not slouches in the brains department – but as far as we know, the combination of a falling weight as a power mechanism, with a mechanical escapement, did not occur until the 12th century AD at the earliest.
The question of who was the first to think of using a coiled spring as a mechanism for powering a clock, will almost certainly never be answered. The invention of spiral springs preceded their use in clocks; they may have first been used by locksmiths, for whom springs were essential for the construction of secure and reliable locks. (Other types of steel springs were also essential to early watchmaking, as well as locksmithing and the development of firearms as well.) The oldest existing spring-powered clock is in the Germanisches Nationalmuseum, in Nuremberg; it was given to Phillip the Good, Duke of Burgundy, in 1430 and it is so complex that it must certainly have had antecedents but the origin of the mainspring seems destined to remain obscure. (Claire Vincent, who for many years curated the watch and clock collection at the Metropolitan Museum of Art, writes in European Clocks and Watches In The Metropolitan Museum of Art, that the likeliest origin for the technology was northern Italy).
“The art of making the mainsprings of watches and clocks is perhaps, of all mechanical manipulations, that which provides the most physical knowledge of the properties of steel.”
– WILLIAM BLAKEY, THE ART OF MAKING WATCH SPRINGS, 1780
Now, the problem with using steel springs to power anything, much less watches and clocks, is that you need the steel to have very specific properties. It has to be both elastic and tough, and it needs to be able to withstand many cycles of use without breaking due to metal fatigue. (A broken mainspring was no mere inconvenience; it could result in the explosive delivery of much of the potential energy of the spring directly into the gear train, badly damaging or destroying the mechanism). The hardness and elasticity of steel is highly dependent on minute variations in its chemistry, and steel alloys are also extremely sensitive, in terms of their final properties, to how they are worked.
The simplest definition of steel is that it is an alloy of iron and carbon. At the low end, you can have as little as 0.002% carbon; at the high end, no more than about 2.14% (this is for simple carbon-iron alloys). More than that, and you have pig iron (which can be further worked into wrought iron or refined further to use as cast iron) and below that, you have almost pure iron which is too soft and malleable to be useful. Within that range, you can produce steels with an enormous variety of properties. Steel objects can be produced with different crystal and chemical properties in the same artifact – swordsmiths in Japan and elsewhere have for many centuries, produced blades with dramatically different metallurgic properties in different parts of the blade.
A mainspring is basically a blade as well – of very thin, very flexible steel, which is coiled on itself into a spiral, and placed in a mainspring barrel. Generally speaking, the spring is attached at the inside to an arbor (a steel rod around which it’s coiled as you wind the watch) and at the outer end, it’s attached to the barrel. The arbor is turned by the crown when you wind the watch, and it’s held in place by what watchmakers call a click (basically a kind of pawl) when the watch is running. Since the arbor can’t rotate, the barrel does, and gear teeth on the barrel engage the pinions of the next gear in the watch.
This next gear is the center wheel of the movement and in a classic watch, the center wheel rotates once per hour. Each wheel in the train rotates more quickly than the one preceding it, so that by the time you get to the fourth wheel (the last one before the escape wheel) you have a rotational speed of once per minute; in a traditionally laid-out watch you get a sub-dial display of the running seconds, by just putting the seconds hand on the fourth wheel pivot. It is an interesting feature of mechanical watches that there is a spiral spring at each extreme end of the gear train – a mainspring at one end, for power, and a balance spring at the other, which acts to regulate the motion of the balance.
One can only imagine the ingenuity and persistence it must have taken to develop the first mainsprings. Despite the fact that we don’t much consider the mainspring nowadays, making them in quantity to a reasonable degree of consistency in quality, was for much of the history of watchmaking, an art in itself and a highly specialized one at that. It is easy to be somewhat mystified by the degree to which watchmaking was compartmentalized almost from the beginning but, if making mainsprings is any indication, it should be obvious in retrospect that this would have to be the case, and that watchmaking would be a collaboration among dozens of different craftsmen.
The manufacture of a mainspring began with high-quality steel, made from iron ores which could differ dramatically in quality; it was then worked, over a number of painstaking steps, in order to produce the finished product and an 18th-century watch mainspring was the product of many days of careful labor. The apparent simplicity of the mainspring to a watch enthusiast of today makes it hard to understand why they were made by specialists, but when you remember just how challenging it is to make a thin blade of very flexible and elastic steel, which can be coiled into a barrel only a few centimeters across, using only manual craft techniques, you start to understand why watchmakers ordered them in rather than trying to make them themselves.
The process is described at length in an 18th-century book entitled, L’Art de Faire les Ressorts de Montres, (The Art Of Making Watch Springs). The treatise was written by an Englishman, working in both England and Holland, named William Blakey, and was published in Amsterdam in French, by Marc-Michel Rey, in 1780. It has been recently translated by Richard Watkins, into English and if, like many of us might, you have dismissed the making of mainsprings out of hand as one of the more menial and less interesting aspects of traditional watchmaking, it is an eye-opener.
Blakey outlines the basic problem at the beginning of his treatise, thus: “The art of making the mainsprings of watches and clocks is perhaps, of all mechanical manipulations, that which provides the most physical knowledge of the properties of steel. By initially discovering the essential qualities to convert iron into steel, the artist cannot fail to recognize, in this work, the various qualities of this metal, such as its hardness, its malleability, its elasticity, etc. To understand what I will say, it is necessary to know that an ordinary watch mainspring is a small, thin blade, from twelve up to twenty-two inches long, bent so that it has the elastic force to make a balance vibrate 540,000 times in thirty hours.”
Making such a mainspring by hand is, to put it mildly, easier said than done. The section of Blakey’s treatise on making mainsprings has a total of 69 individual sections, and details dozens of operations, from which type of steel to use (English, he felt, was the best, with German steel shipped through Danzig deemed a close second) to drawing out the steel into wire, followed by many additional steps which included using special filing jigs to give the mainspring a very fine taper from one end to the other. If all went well, what you had at the end was a spring which could, thanks to the precision of its construction, unwind in a mainspring barrel without any of the coils rubbing against each other, and which, when used in conjunction with a fusee, would give a running time of thirty hours. Blakey learned his art from his father, who was responsible for many innovations and improvements in mainspring manufacturing and his work seems a reasonable representation of the gold standard for mainspring manufacture in the late 18th century. It was, he noted, sufficiently difficult to make good mainsprings that watchmakers often paid handsomely for good ones from suppliers. Blakey remarks, for instance, that in his father’s day, ” … the watchmakers of Paris usually bought their springs from Geneva, and paid triple and quadruple for those which they got from England.”
Making steel with good quality control standards, and in large amounts, is obviously essential to industrial watchmaking in general, and mass production of mainsprings in particular, and yet it didn’t really take off until almost a hundred years after Blakey wrote his treatise. By then there had been myriad other innovations that made industrial watchmaking possible, including the first true milling machines; these meant that for the first time, factories could produce large numbers of interchangeable components. The first inexpensive process for making steel in industrial quantities was the Bessemer process, in which air is forced through molten iron under pressure, burning away impurities. Henry Bessemer, the inventor, began using the process which bears his name at his steelworks in Sheffield, England, in 1855 and was granted a patent for the process a year later – it was so useful that the last Bessemer converter remained in service until the 1960s.
Carbon steel mainsprings were a metallurgical wonder in their time but even at their best, they still had flaws. Other than breakage, the biggest problem was the loss of elasticity over time – the mainspring would eventually become weak enough to negatively impact both power reserve overall, and balance amplitude during the running period of the watch. The only option at that point was to replace it. After World War II, carbon steel mainsprings were gradually replaced by metallurgically more complex and more sophisticated alloys, which did not suffer from the problem of “setting” (losing their elasticity) to nearly the same degree, much less breaking outright. Today, making mainsprings is perhaps even more highly specialized work than it was prior to the 20th century, and thanks to the need for absolute consistency in production, much of the work is now automated. It is however no less interesting, and still far more complex a process than you might imagine – Peter Speake-Marin has an excellent look at the basics of today’s processes for mainspring production, at The Naked Watchmaker.
One such modern mainspring alloy is Nivaflex, made by Nivarox, which is a highly complex material. In an article on modern mainsprings, Gisbert Brunner notes: “By weight, Nivaflex consists of 45 percent cobalt, 21 percent nickel, 18 percent chrome, five percent iron, four percent tungsten, four percent molybdenum, one percent titanium and 0.2 percent beryllium; carbon accounts for less than 0.1 percent of this alloy’s weight. Increasing the percentage of beryllium in an alloy further increases its strength and hardness, factors that are important for miniaturization.”
A far cry, this, from the hand-crafted, blued carbon steel mainsprings of yesteryear. Another modern mainspring alloy is SPRON 510, which is manufactured and used by Seiko Instruments Inc. (SII). SPRON 510 is an alloy of cobalt, nickel, molybdenum and other elements, and like Nivaflex, it is highly resistant to breakage and to strain-induced fatigue. It is non-magnetic and offers very even power delivery over the running time of the watch. Modern mainspring alloys, as well as modern high-precision fabrication of other gear train components, have made it possible to make wearable watches with longer and longer power reserves. The Lange 31, for instance, which debuted in 2007, has a full one-month running time courtesy two extremely long (185 centimeters) mainsprings. It is interesting to compare the Lange 31 to its compatriot, the 16th-century German portable clock pictured earlier in this article – they are roughly comparable in size but it is likely that the clock (or watch – it’s certainly small enough to be considered a candidate for the term) ran for, at best, a day.
We said at the outset that without mainsprings, there is no watchmaking. As mainsprings are indispensable for portable mechanical timekeepers, this raises the question: when was the first watch made? Like the origin of the mainspring, this is unlikely ever to be known with any certainty. Too much time has passed; record-keeping was spotty at best in the 15th century, and in any event, there is no clear distinction that can be made between a small portable clock and a watch per se, though as horologist Kenneth Ulyett remarked of the origin of the lever escapement, were the matter certain, ” … many keen horologists would be deprived of the pleasure of arguing with each other.”
Mainspring technology continues to advance although, as with many other aspects of mechanical horology, improvements tend to be incremental rather than revolutionary. Mainspring production nowadays, thanks to the precision necessary in modern watchmaking, as well as the complexity of the alloys and metallurgy used, is largely automated. There are however occasionally signs that more dramatic advances might be possible.
For a time in the mid-2000s, Cartier invested heavily in experimental movement technology, and the two most spectacular concept watches it produced were the ID One and ID Two timepieces. ID Two had four mainspring barrels, with a most unusual mainspring material. Instead of using carbon steel, or a high-tech alloy, ID Two used fiberglass springs and this in combination with many other innovations in the construction of the watch, gave it a power reserve of 32 days. Unfortunately for fans of advanced watchmaking as a form of intellectual entertainment, Cartier seems to have decided not to pursue the technology, but it remains a fascinating what-if in modern technical horology (for more, check out Ben Clymer’s In-Depth post from 2012).
The complexity of modern mainsprings gives one all the more reason to admire the skill of craftsmen of the past, who had to work with knowledge handed down from master to apprentice (or in the case of William Blakey, father to son). The manual skills necessary to make a mainspring, starting with only raw steel wire of uncertain quality, required years to learn, and more years to perfect, and raised mainspring making to the level of a high craft. These skills and methods are largely lost. Like the whole history of watchmaking, the history of mainspring making is as much a story of advances in materials science and automation as it is anything else. But though it labors largely unseen, unconsidered, and certainly largely unheralded, the mainspring, if you consider it closely, embodies in its evolution the past, present, and future of watchmaking.
Notes and further reading:
Read William Blakey’s treatise on hand-manufacturing mainsprings in the 1780s here, in Richard Watkin’s 2014 translation. Other horological texts and translations by Watkins can be found here. For an excellent introduction to modern mainspring theory, see Gisbert Brunner’s 2015 article for Watchtime here. Finally, for a rare look inside the doors of a modern mainspring manufacturer, see “Making Mainsprings At Générale Ressorts” by Peter Speake-Marin, over at The Naked Watchmaker. Headline image, SPRON 510 balance spring, with mainspring barrel, balance, escape wheel and other components; for more technical information on SPRON 510, visit SII online.