You might say, if you were a casual (or, for that matter, non-casual) observer of the watch enthusiast scene nowadays, that mechanics don’t get no respect. The vast majority of the discussion about watches focuses on cosmetics – I don’t mean that in a negative way; the reality is that most watch purchases and most of the subsequent enjoyment focuses on elements of a watch that don’t have anything to do with how it performs as a watch. Dial colors, case and bracelet design, quality of case finishing and of dial furniture, the degree to which hand work is involved, and even movement finishing to a certain extent, all have much to do with how a watch looks and little if anything (with a caveat for movement finishing, which as it is applied to working surfaces, has a major influence on performance) with how well the watch tells time.
We have long taken accuracy for granted, and it surprises no one to have a mechanical watch that can keep time to within a few seconds a day or better, and yet how we got there is a long, complex journey – one which has about it a bit of the air of the inevitable.
What Time Is It, Anyway?
At the core of the conversation is the notion of accuracy. This is the degree to which the watch agrees with an external, more accurate time standard. Not long ago, a HODINKEE reader asked a quite interesting question which we discussed a bit in Episode 95 of HODINKEE Radio, which was, how did watch and clockmakers know, historically, whether a particular clock or watch, or a particular aspect of the mechanism, was an improvement in accuracy if the standard for accuracy is an existing clock or watch? The answer is that, for wristwatches, time standards have usually been clocks. Clocks were accurate long before watches. This is thanks to the invention of the pendulum which, by the beginning of the 18th century, was already capable of incredible accuracy – a clock by Tompion, finished in 1676 and installed at the Royal Observatory at Greenwich Park, just outside of London, was so accurate that the then-Astronomer Royal, John Flamsteed, was able to use it to determine that the speed of the Earth’s rotation was constant.
Such clocks would have presented an aspirational challenge to watchmakers, who could only dream of such precision. Pendulum clocks can be as precise as they are for several reasons, but among the most important of these are that they are generally stationary and protected to some degree from the elements. Moreover, the precision of an oscillator, such as a pendulum or balance, in a clock or watch depends very much on the restoring force – the force that tends to bring the oscillator back to its neutral position – being proportional to the driving force. In a pendulum, this is easily done, as gravity tugs the pendulum back to its neutral position more or less with the same force with which you push it (as anyone who has ever taken a swing to the chest pushing a child on one in a playground can attest).
Clocks, in turn, could be set against astronomical phenomena. Particularly of use as a time standard were the transits of stars across fixed points in the sky; these transits occur with extreme regularity and predictability and were the final time standard against which clocks were measured, right up until the 20th century, when, first, quartz oscillators and then atomic clocks, set a standard in which a clock might have better than one second precision over the known age of the entire universe.
Clock Vs. Watch
Clocks, therefore, can be made so accurate that they vastly exceed any need for precision which anyone might need in daily life, and they had already done so more than two and a half centuries ago. So why did watches lag behind so badly?
The answer is, again, pretty simple. A watch, first and foremost, is portable (it had better be, or it’s not much of a watch), and wristwatches especially are subjected to an awful lot of abuse, including physical shocks, temperature changes, occasional but potentially disastrous exposure to magnetic fields, and on and on. As watchmakers recognized early on that their clients were not likely to become more careful, watches needed to become better and better at resisting external disturbances. The other problem is that a watch does not have a pendulum. (A pendulum in a watch, you would think, is not such a hot idea, and you would be right, but that hasn’t stopped some people from trying. Pendulums were, for many years, so strongly associated with accuracy that some makers of cheaper pocket watches would actually put dummy pendulums in them which were visible through an aperture in the dial, which I would think would actually be less reassuring.)
Instead, a watch uses a circular balance, and fans of modern watches might be surprised to find out just how long those have been around. The balance actually pre-dates the pendulum, in one form or another – balances, and a related device called a foliot, which is essentially a bar mounted on a pivot and rotating in the horizontal plane, date back to the very earliest known mechanical clocks, which were probably developed towards the end of the 13th century (record keeping back then was terrible and often interrupted by war, pestilence, illiteracy, and what have you – on top of that, the earliest known clock movements in Europe were iron, and the centuries have reduced most of them to rust). The balance, however, in its earliest form did not have a natural harmonic frequency – instead, the escapement in these early clocks, called the verge, simply knocked the balance or foliot back and forth with no real regulating force to steady its rate, like a bored child whacking a rubber ball back and forth between its hands. As a result, you were lucky to get an hour-per-day accuracy out of the things, although this didn’t really matter as you were lucky to get thirteen or fourteen hours’ running time out of them as well.
Things chugged along in this fashion, in watchmaking, for a couple of centuries until the mid-1600s, when the balance spring was finally applied to the balance. The reason this was so critical is because the balance spring provided, like gravity, a restoring force proportional to the driving force (at long last), and in a single stroke, accuracy went from an hour a day to minutes per day – suddenly, smaller clocks and, more importantly for our purposes, watches could be thought of as potentially serious timekeepers. A restoring force proportional to a driving force is critical in a clock but even more so in a watch, as it means the balance will maintain a stable frequency irrespective of the energy in the mainspring (that’s the idea, anyway – although, as we will see, there is truth in that old saying about the best-laid plans of mice and men).
This is a property called isochronism. Isochronism is not a given just because you throw in a balance spring – the number of coils, materials used, where the inner and outer attachment points are, were all things that had to be worked out empirically and with a lot of long, slow, painstaking work (which continues to this day). But, in principle, the single greatest problem in watchmaking had been solved (and this was quite a while ago; about 1657, thanks to the research of the Dutch scientist Christiaan Huygens and the English researcher Robert Hooke).
However, as is so often the case with knotty problems, cut off one head and three more grow in its place. With the invention of the balance spring came a whole host of new and extremely intractable problems.
The Pursuit Of High Precision
Throughout the history of watch and clockmaking, one theme constantly recurs – the more precise a timekeeper gets, the more you have to start taking into account things you could ignore before. Take the balance spring. Once you start getting minutes-per-day accuracy and better precision, suddenly you start to notice things that you hadn’t. Temperature, for instance, starts to become a factor, and you have to figure out what to do about the fact that due to the thermal expansion of the balance, and the changes you get in hairspring elasticity as temperature changes, rate changes unpredictably as temperature changes – you need to figure out some form of temperature compensation. Early efforts at temperature compensation saw an awful lot of different approaches, but eventually, by the early 1800s, the standard approach was to use a balance made of a layer of two metals – brass and steel, for instance – that expanded and contracted at different rates. The balance is still a circle, but it has cuts in the rim so its diameter can expand and contract as temperature changes, which offsets the effects of temperature on the balance and spring – not perfectly, but a hell of a lot better than nothing.
Gradually, two basic directions emerged in watchmaking over the course of the 19th century. On the one hand, you had watches which were designed with a view to changes in style, and the primary goal was to make watches as elegantly thin as possible. This tendency reached its pinnacle at the end of the 19th and beginning of the 20th centuries, when the so-called “knife” pocket watches, some of which had movements so thin it would be difficult to reproduce them today, were made. While these were made with a great deal of attention and care – ultra-thin watchmaking is extremely demanding, and very few watchmakers ever made true ultra-thin movements – such watches, as a rule, do not have precision and accuracy as their primary goal.
The other direction, however, was oriented towards precision as a first priority. Such watches tended to have thicker, more rigid movements; they also had large, high-mass balances to ensure a more stable rate; they invariably incorporated cut bimetallic balances for temperature compensation, and they were adjusted to at least five or six positions to ensure a minimum variation in rate due to the effects of gravity. As a balance spring was thought to “breathe” more naturally if there was no regulator index (which can be used to change the effective length of the balance spring, thereby changing the rate of the watch), these were often freesprung as well, with a Breguet overcoil to further reduce rate variation across positions. Often, a Maltese cross stopworks would be incorporated, which restricts power output from the mainspring to only that segment with the most even delivery of torque (or, in the case of English watchmaking, a fusée and chain).
Really top-level precision obsessives might wish for a watch with a chronometer detent, rather than a lever escapement, although lever watches achieved extremes of precision as well and offered much better resistance to shocks. At the end of the 19th century, then, this was the classic form of the high-grade, chronometer pocket watch: generally, time only, with a large, freesprung adjustable-mass balance with temperature compensation, an overcoil balance spring, either a Maltese cross stopworks or a fusée and chain, and either a high-quality lever or chronometer detent escapement depending on the preferences and, I presume, the friskiness of the potential owner.
Such watches would have been made on a very artisanal basis, one at a time, with laborious adjustments for isochronism, as little variation among positions as possible, and temperature (and often the movement would be engraved to this effect, with some variation on, “Adjusted to heat, cold, isochronism, and 6 positions.”). However, it is also true that mass-produced watchmaking was hard on the heels of these chronometrically oriented superwatches and with American watchmaking in the lead. American watchmakers did not make especially high-craft watches as a rule, but they were robust, accurate, reliable, and above all, cheap (relatively speaking). The U.S. watch manufacturers were of enormous concern to the Swiss as they threatened to simply out-produce the Swiss in terms of numbers, at least perform equally in terms of accuracy, and just as importantly, significantly undercut them in costs. And it was in the early 20th century that the stage was set for true high-precision on a mass-production basis.
The 20th Century: Materials Matter
For most of the history of watchmaking, the basic tools and materials had changed very little: Brass, either gilt or rhodium-plated for plates and bridges (or, in some instances, German silver which is arguably a kind of brass; brass is an alloy of copper and zinc, while German silver is copper and nickel, sometimes with zinc as well), steel of course, and … and, really, not much else; an occasional use of gold for working parts; rubies (first natural and then synthetic) as very low friction bearings, and lubricants, which were originally animal and vegetable greases, but which were gradually replaced by longer-lasting synthetics. It is interesting to think what alchemy was achieved with these rather intractable materials wrung, as you might say, from the bowels of a resisting Earth. With the accumulated knowledge of centuries (just making good mainsprings was a high craft passed down from father to son at one point), makers could create watches and portable timekeepers like marine chronometers on which everything from getting to a train station on time, to the fate of the fleets of mighty nations, might depend.
As the 20th century’s first decades went by, a quiet revolution occurred in watchmaking. This was the invention of nickel-steel alloys with very low coefficients of expansion – in other words, their physical characteristics changed dramatically less with temperature changes than conventional materials like brass and steel. The fact that these alloys have a temperature coefficient of almost zero changed watchmaking forever. Invar (the name comes from “invariable”) and Elinvar (“invariable elasticity”) were alloys whose properties were researched extensively by the Swiss physicist Charles Édouard Guillaume, who won the Nobel Prize for his work in 1920.
Invar became the gold standard for pendulum rods (there is a nice Easter egg for horology fans in the Terry Pratchett novel, The Thief Of Time, in which a woman attempting to commission the most accurate clock ever made finally gets the clockmaker’s attention by saying silkily, “Do you need Invar? I can get you any quantity of Invar.”). Elinvar balance springs, plus solid balances made of beryllium bronze alloys, gradually became the standard. Automatic winding systems, plus new mainspring alloys with very flat torque curves over most of their power reserve, did away with the cut balance, the plain steel overcoil balance, elaborate stop works and fusées, and all the centuries of accumulated knowledge on how to make them work together in an improbable but beautiful symphony of accuracy. The lost horological decorative arts get a lot of attention these days, but almost invisibly, the advent of high tech alloys swept an entire world of technical skill and knowledge irrevocably into the past as well.
Materials: There’s A New Sheriff In Town
These developments in materials science meant that, once again, the precision watch underwent a metamorphosis. A precision chronometer-grade wristwatch would now tick these boxes: still a freesprung, adjustable mass balance, with overcoil balance spring, but the balance would be made of a beryllium-bronze alloy like Glucydur, and with the spring made of one of the nickel-steel alloys like Nivarox. The watch is automatic, with a slipping bridle mainspring made of a material like Nivaflex, which has an extremely complex composition: 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, with just a smidge of carbon (less than 0.1 percent) thrown in for God knows why … to appease the Old Gods Of Carbon, maybe.
The lubricants are modern synthetics, and, thanks to the advent of modern multi-axis computer-controlled machine tools, the watch in question is going to be exactly like its siblings, which can number in the hundreds of thousands with no more to tell them apart than one Clone Trooper from another. I hasten to add that this is a good thing and a long-sought goal of horology – inconsistency in production is a gremlin to be rooted out and exterminated by any means necessary.
One word on escapements: with a couple of exceptions, the lever escapement rules the roost and has for many decades. There are several reasons for this. The first and most important is simply that it works. It works extremely well – it is inherently resistant to shock, its properties have been researched and refined extensively since Mudge made the first lever watch in 1755, and moreover, it is capable of keeping really excellent time, with the best examples (from, for instance, companies like Rolex, whose Chronergy escapement has a proprietary geometry) perfectly capable of keeping time to within a maximum daily variation in rate of just a couple of seconds a day. To date, the only other escapement to be produced in really industrial quantities is the co-axial escapement, and in Omega’s watches, in combination with further innovations in materials science, including silicon balance springs, escape wheels, and pallets, both magnetism and degradation of lubricants are becoming non-issues – certainly over time spans of less than a decade.
One of the first distinctions you learn about when you get interested in precision timekeeping is the distinction between precision and accuracy. Precision is more or less synonymous with rate stability – the property of an oscillator, that its frequency should not vary, which by now we can see is harder than it looks. Pendulum clocks go in one direction – that of isolation; the most precise can have their rate disturbed by something like the tidal forces of the Moon passing overhead, and they live in vibration-silencing, temperature-controlled underground vaults. Wristwatches go in another – that of arming themselves against the slings and arrows of outrageous fortune, and developing materials and designs which allow them to pass, like Odysseus lashed to his mast, between the threats of physical disturbances and magnetic disruption unscathed. While there is obvious overlap between the two fields – a watchmaker never met a bar of Invar they didn’t like – each has its own fascinatingly unique characteristics.
Precision is of interest because, without it, real long term accuracy cannot be had. You can be precise without being accurate – a chronometer which gains ten seconds a day on its rate, no more, no less, is actually extremely precise in that its rate does not vary, and you can easily derive accurate time by subtraction. An imprecise watch, which wanders from day to day on its rate, may seem for days at a time to be precise if you happen to get lucky with rate variations canceling each other, but sooner or later, luck, which has no place in precision horology or precision anything else, runs out, and you are stuck looking at the back of a departing train or a closed departure gate.
That it is possible for mechanical watches to be produced in the millions and still keep time to within chronometer specs, or even significantly better, is something that would have left the masters of yore, who labored for months and years in isolation on single high-precision timepieces, awed. However, that is not to say that they would not have understood how it was achieved. It’s wonderful to think of what the reaction of someone like Breguet would be if you showed them a modern watch from a company like Rolex. Breguet would certainly be struck by many things about the watch, and he would, for sure, have no end of questions to ask, but at the same time, there is probably nothing in the watch which would have mystified him from a basic technical perspective. By the time of his death, everything that makes a Rolex or other modern watch tick was already firmly in place, and the accuracy which we now enjoy in good watches, and take for granted, is based on a quiet, relentless quest for refinement in technology which is three centuries old. Fine watchmaking firms often make much of their watches as a bridge to the past, but that assertion is firmly grounded in reality, and what machines and computers achieve today by the millions would not exist were it not for the individual minds and hands, infinite patience, and probing curiosity of the masters of yesteryear.