The minute repeater is both the most sophisticated of complications and the most atavistic — sophisticated, in that it requires enormous care and skill to make one that works at all, much less one that works well and sounds good; atavistic, in that it brings us back to the very earliest attempts to make mechanical timekeepers. The first clocks weren’t made to be seen, but rather, heard, and when entire communities relied on clocks to tell the time, tolling the time was preferable.
It’s easy for us today to forget just how dark mankind’s nights were, in the days before ubiquitous electric light, but there must have been something very reassuring about hearing the time tolled on a bell — or, when the first chiming watches were made, on a mechanism that you could keep on your person. The repeater today no longer serves as a psychological hedge against the terrors of the night, nor is it useful in the same way that it once was, now that artificial light (and luminous watch hands, to say nothing of electronic watches that simply make their own light) has made it possible to see the time whenever one wishes. But the fascination of its mechanism still remains, and the childlike pleasure it offers is all the more rich for depending on one of the most complicated mechanisms in watchmaking.
The basics: a repeating watch is one that chimes the time on demand, rather than in passing (the grandfather or mantel clock does the latter, singing a song of the hours as each one expires). In most cases nowadays, the repeater is set chiming by pushing a slide set into the side of the case, though other methods exist and are still in some cases in use, including pushers rather than sliding levers.
The purpose of pushing the slide is twofold: to activate the chiming mechanism and also to provide the power necessary for the gongs to sound. When one pushes the slide, one winds up a second mainspring inside the watch. The energy for the going train — that part of the watch that carries torque to the escapement and balance wheel, which actually keeps time — comes from a spiral spring housed in a drum-shaped mainspring barrel. Operating the chiming mechanism takes up so much energy that to use the mainspring would run the watch down very quickly (and the loss of torque would tend to disrupt timekeeping as well). The solution, therefore, is to have a second mainspring, in its own barrel, to power the chiming mechanism.
Pushing the slide home, and then releasing it, lets the repeater go to work. The repeating mechanism is basically a tool for evaluating the internal state of a watch, and translating that information into chimes. A typical repeater has two hardened steel gongs, and first chimes the hours on the lower of the two, then the quarter hours, as a double stroke (ding-dong) on both gongs for each quarter hour, and then finally a stroke on the higher-pitched gong for each of the minutes. The maximum number of strokes a repeater will strike, then, occurs at 12:59, when a repeater strikes 12 hours, three-quarter hours, and 14 minutes (those wishing to test a repeater’s functioning can easily do so by setting the watch to 12:59 to see if the repeater sounds the correct time).
The repeater works as follows. The hours, quarter hours and minutes are each “read” off so-called snails — circles of metal with steps of varying depth cut into them, which rotate along with, and are driven by, the gears of the going train. The hour snail has 12 steps, and is indexed — turned — 1/12th of its circumference each hour by a star-wheel gear. The quarter snail has four steps, and the minute snail has the most exotic shape of all. Remember that the repeater chimes the minutes from none to 14, and that there are four possible sets of 14-minute intervals per hour (plus four one-minute periods when no minutes should be struck). The minute snail is, therefore, shaped like a four-armed starfish, with curved limbs; each limb or arm has 15 levels (14 steps plus an uppermost step which, when the minute rack contacts it, prevents any minutes from striking).
To sense the position of each snail, there are three racks — complicated levers, with a metal feeler on one end and ridged teeth along the opposite edge. The number of teeth on each rack corresponds to the number of steps on each snail. For instance, up to 12 hours may be struck; the hour snail therefore has 12 steps and the hour rack has 12 teeth.
When the repeater is activated, the mainspring is wound and each rack is released to fall onto the step of its associated snail — the hour rack onto the hour snail, quarter rack onto the quarter snail, and minute rack onto the minute star. How far each rack falls depends on the depth of the step onto which it falls. The more chimes to be rung, the deeper the step; the step on the quarter snail for three quarters of an hour is deeper than that for one. As each rack falls, its teeth pass the projecting trips of the hammers that will sound the time (the quarter rack has two sets of teeth, to activate the higher- and lower-pitched gongs in succession for each quarter of an hour). The teeth of the racks and the hammer trips are arranged so that when the repeater is being armed, the rack teeth can pass the trips without causing the gongs to sound.
Once the racks have fallen onto their respective snails and the slide has been released, the watch is ready to chime. The mainspring for the repeater first drives the hour rack back to its resting position, and as it does, the teeth of the hour rack flick the trip for the hour hammer, ringing the hour gong — the number of times determined by how far the hour rack has been able to fall on its snail; the deeper the step, the greater the number of hours. The same principle applies for the quarter rack and for the minute rack.
The hour snail is mounted on a 12-toothed star wheel, which is indexed (pushed forward one tooth) once per hour by the cannon pinion. The cannon pinion is a hollow pipe that snaps onto the axis of the center wheel of the going train of the watch (a refresher: the going train consists of the mainspring barrel, center and third wheel, fourth wheel — which turns once per minute, and to which the seconds hand is attached in a traditional, sub-seconds layout pocket or wrist watch — and escape wheel, which gives impulse, in a lever escapement watch, to the lever, which in turn impulses the balance wheel). Since the center wheel, and thus the cannon pinion, turns once per hour, its rotation forms the basis of controlling the rotation of the repeater snails on their axes — each snail turns once per hour and each has a number of steps appropriate to how many chimes it can ring (four for the quarter snail, 12 for the hour snail, and four arms with 15 steps each for the minute star).
The hour snail is not directly attached to the cannon pinion, but the quarter snail and minute snail are, and this presents a problem. Since both the quarter snail and minute snail are attached to the cannon pinion, it is possible for the tail of the quarter or minute rack to fall onto the lip of the highest step of their respective snails and jam the entire watch (since the cannon pinion is friction fitted onto the center wheel of the going train). This would obviously be a disaster. The moment when this could potentially occur is at the beginning of the hour, in the case of the quarter snail, and at the beginning of each quarter hour, in the case of the minute star. A related problem, which could occur at these moments, is that the tail of the rack might accidentally drop onto the lowest step when it should drop onto the highest step. If this were to occur in the case of the quarter rack, three quarters would chime when none should chime (just after the turn of the hour), or 14 minutes would chime when none should chime (during the first minute after each quarter).
To address both problems, a special part is hidden underneath both the minute snail and the quarter snail: the so-called “surprise piece”. Understanding the function of the surprise piece has perplexed many amateur and professional horologists alike (a manual on quarter-repeater repair by François Crespe and translated by Richard P. Watkins, dating from 1804, and written in the form of a dialogue between a master watchmaker and an apprentice, has the latter ask at one point, rather plaintively, “Please give me a description of the surprise piece, which few horologists can explain?”). The surprise pieces are confusing because, unlike most parts which are named for their function, they are named for how they look in action — they are essentially extensions of the highest step of their respective snails, which are normally held out of view, but which snap into view just an instant before they are needed… surprise!
Basically each surprise piece is an extension of the uppermost step of its respective snail. They act as a form of insurance. Since each extends the uppermost step, they ensure that the tail of the rack (either the quarter or the minute rack) can’t fall onto the bottom step and chime quarters, or minutes, when none should be struck. They also ensure that the tail of the rack will fall properly onto the highest step (extended by the surprise piece) rather than jam against the tip. In order to keep them from blocking the watch from chiming the quarters or minutes when it should, once they’ve done their job, the surprise pieces are pushed back under their respective snails until they’re needed again, clearing the way for the longer minutes and quarters to chime.
The second major problem with a repeater has to do with what happens if the slide arming the repeater mainspring isn’t pushed all the way down. Under such circumstances, it might be possible for the repeater to chime the time incorrectly, as the barrel would not rotate fully, causing too short a chime.
To prevent this from happening, another part unique to repeaters exists — the “all-or-nothing piece”. This is basically a spring lever which blocks the repeater works from operating unless the arming slide is pushed all the way down. If, and only if, the slide is pushed fully home, the all-or-nothing piece shifts out of the way, unblocking the repeater train and allowing the watch to chime the correct time.
Yet another exotically named component is the spring-like part that flicks the minute snail surprise piece into place. This part is in English watchmaking literature called a “flirt”. The flirt, in order to keep it from rubbing unnecessarily across one of the four tips of the starfish-shaped minute-snail surprise piece four times an hour, is kept clear of the surprise piece by an isolator, which only lets the flirt drop into place and flick the minute-snail surprise piece forward when the slide is actually pushed home.
The minute repeater isn’t the only variety of repeating watch; there is the quarter repeater (which only chimes the hours and quarters), the half-quarter repeater (which chimes the hours, the quarters, and then an additional strike if it’s more than seven-and-a-half minutes past the hour) and the five-minute repeater (which strikes hours, quarters and the number of five-minute intervals past the quarter) — and each preceded the invention of the minute repeater itself.
The most sophisticated chiming watch is the minute repeater with grande et petite sonnerie. This is a watch that can chime the time on demand, but also chimes the hours and quarters en passant, without the owner having to press a button or operate a slide. A grande et petite sonnerie can be set to chime only the quarters past the hour, both the hour and the quarters past the hour, or it can be set to silent mode, in which case no chiming en passant occurs, and the watch will only chime (as a minute repeater) if asked. Because of the very large additional power requirements, the grande et petite sonnerie has a very large separate mainspring barrel — which can easily be larger than the mainspring for the going train — to power the strike.
It’s probably clear even from this cursory explanation that repeaters are not to be trifled with, either by careless owners or by unprepared watchmakers. Vintage pocket repeaters offer many charms to collectors (not the least of which are much lower prices — usually — than new wristwatch repeaters), but in many cases, they require extremely expensive and time-consuming service to work properly — made even more difficult thanks to the fact that so many have been the victims of botched service attempts in the past by overenthusiastic but unqualified technicians. If it’s necessary to replace vintage repeater parts, the difficulty of repair skyrockets (along with the time and cost), as vintage repeater parts in general aren’t interchangeable (and in any case, many manufacturers are defunct) and the watchmaker will have to make one from scratch, right down to finishing it to a level appropriate to the rest of the movement. Making and adjusting a repeater to work properly is one of watchmaking’s biggest challenges — each of the many additional parts of the repeater train have to be synchronized perfectly so that each event in the chiming of the time happens when it should, and that’s before you even start to consider how to control the quality of the sound itself.
In considering the challenges a repeater poses to a watchmaker, it’s important to first understand the fact that as a complication, the repeater is incredibly unforgiving when it comes to precision of construction. Those of us lucky enough to examine a minute snail under a loupe, for instance, are astonished at both how perfectly formed its curves must be, as well as how incredibly minute each step is — the tips of each arm are especially delicate. We had an opportunity last year to see a vintage ultra-thin Cartier pocket repeater being restored at the workshop in La Chaux-de-Fonds — the tip had broken off from one arm of the minute snail — and Cartier’s restoration team was remaking the part. The original star had been made entirely by hand and was a work of art in itself, which must have taken many dozens of hours to fabricate — and it was just one of the dozens of repeater-train components in the watch. The slightest deficiency in either accuracy of construction or care in assembly will result in a watch that fails to work properly, and many repeater components are so delicate that they must be handled with the greatest care and unflagging concentration to avoid damaging them.
One of the most critical aspects of the repeater is how fast it chimes. According to Patek Philippe’s Thierry Stern, there are basic standards but also much room for personal taste (Mr Stern, who still validates every Patek repeater before it leaves the workshop, as did his father Philippe Stern, has told us that different markets seem to prefer slightly faster or slower tempi as well). The speed of chiming is controlled by a gear train driven by the rotating mainspring barrel of the repeater train — since the speed of rotation of the repeater mainspring controls the rate of chiming, controlling that speed of rotation will set the tempo of the chimes. At the end of this gear train is the governor.
Over the centuries, several different types of governors have been developed. Today the two most common varieties are the anchor governor and the so-called fly, or centrifugal governor. The former is the more traditional (and older) design; it works basically the same way the anchor escapement in a lever watch works. At the end of the governor train, there’s an escape wheel with an anchor whose pallets alternately lock and unlock the escape wheel. The rate at which this occurs controls the speed of rotation of the repeater mainspring and hence the rate of chiming. The fly governor is based on the same principle as the centrifugal regulator invented by James Watt for steam engines — fitted onto the last wheel in the governor train is a two-armed spindle with curved arms attached to it; these arms are weighted at the end. As the spindle rotates, the curved arms bend outward (controlled by two extremely small springs) — how far out they are allowed to go controls the speed of chiming. Each of these systems can be adjusted to make the chimes ring faster or slower, and in recent years, there has been some interesting innovation in this aspect of repeaters — Vacheron Constantin, for instance, has patented a regulator that uses centripetal force to close a pair of arms around a central post; and Seiko, for its Credor Sonnerie and Credor Repeater, uses a unique system of air vanes, which, as they spin and encounter air resistance, slow the speed of chiming to a pleasing tempo. The advantage in both of the latter is that they are essentially silent (and in the case of the Seiko chiming watches, there are no extraneous mechanical sounds at all, as both watches are powered by Seiko’s Spring Drive movements, which produce neither tick nor tock).
IN CONSIDERING THE CHALLENGES A REPEATER POSES TO A WATCHMAKER, IT’S IMPORTANT TO FIRST UNDERSTAND THE FACT THAT AS A COMPLICATION, THE REPEATER IS INCREDIBLY UNFORGIVING WHEN IT COMES TO PRECISION OF CONSTRUCTION
We were recently fortunate enough to visit Vacheron Constantin’s manufactory in Geneva, where a master watchmaker who assembles and adjusts repeaters made by this most venerable manufacturer was kind enough to explain to us in detail some of the basic factors that affect the sound of a repeater.
Since a repeater is, in addition to being a watch, also a musical percussion instrument, many of the same considerations apply to it that would apply to any other musical instrument. In addition to tempo, the clarity, volume and tone of the chimes are all elements that can be, and are, individually tweaked to ensure a pleasant listening experience.
As the vocal chords of the diva that is the repeater, the gongs are obviously essential to a good sound. Repeater gongs are made of hardened steel and their exact composition is usually a closely guarded secret. (Once during a factory visit, when we asked what kind of steel the gongs of a repeater were made of, our host blandly replied, “An alloy.” It was clear that the subject was closed.) The gongs must be firmly affixed to the movement plate; loose gongs make for a bad tone. The amplitude at which the gongs vibrate when struck by the hammers can be controlled by filing away minute bits of steel near the gong’s base — leaving too much metal will produce a chime that is unpleasantly metallic and which also dies away too quickly. Removing the correct amount of metal helps the tone of the gongs become rounder and richer, diminishes harsh metallic overtones, and allows the gong to vibrate longer, producing a pleasing delay. Altering its length can also control the pitch of the gong; tiny amounts of steel are filed from the tips of the gongs until the correct pitch has been reached.
Needless to say, since metal once removed can’t be added back, watchmakers performing these adjustments may be presumed to do so with a certain level of feeling.
Another critical parameter is the physical relationship of the movement to the case. The more exactly the movement fits the case, the more sound energy is transmitted to the case, which, since it acts as a resonator to amplify the sound of the gongs, will produce a louder sound the more energy reaches it. Vacheron Constantin matches its repeaters to specific cases that are not interchangeable due to the very exact fit between case and movement; it also uses a screw fixed to the foot of the gongs and touching the caseback in order to better transmit vibrations to the case.
Case material (and construction) is yet another important element. A watchcase has as its first task the protection of the movement, but in a repeater it must also transmit and amplify sound. The best material for a repeater case is stiff and light; excessive density will tend to muffle the sound. Platinum, it’s generally agreed, may be the most noble of luxury metals, but it’s so dense that it tends to muffle the sound of a repeater (though as we’ll shortly see, repeaters exist to defy generalizations). Rose gold is the traditionalist’s favorite and is said to give an especially pleasing warmth to the tone, but in search of better performance, modern repeaters can be had cased in steel, titanium or even carbon fiber.
Finally, the relationship of the hammers to the gongs must be adjusted for the best sound. The strength of the strike, as well as the amount the hammers are allowed to recoil, must both be carefully adjusted. The gongs must be struck with enough energy for a good volume, but hit them too hard, or allow the hammers to stop too close to the gongs, and the vibration of the gongs will actually make them strike the hammers, producing unpleasant extraneous noise and deadening the sound as well. At the same time, the stops for the hammers mustn’t hold them too far from the gongs when at rest, or too much energy will be lost when the hammers fall, and volume will suffer.
With all these variables, making a repeater with a good tone, good volume and pleasing tempo is clearly one of the biggest challenges in watchmaking, and the time it takes to do so can be measured not in days, but (depending on the complexity of the watch) weeks or even months. It’s not possible to produce a musically superior repeater industrially — since the quality of a repeater is judged subjectively, each has to be adjusted and tuned by a specialist, and quality can vary dramatically (we have heard repeaters that sound almost eerily ethereal, and unfortunately also heard others — in some cases where quality of tone had clearly been sacrificed to brute force volume — that were unpleasantly metallic and obviously poorly regulated). In no other complication, perhaps, is the user of the watch in such intimate communication with the watchmaker.
And with all this, repeaters remain mysterious. Thierry Stern of Patek Philippe recounted to us, some time ago, a story that’s stayed with us. “We had a repeater to test once,” he said, “that should not have been one of the very best in tone — it was in a platinum case, and we make very good repeaters in platinum cases; but of course, it’s more of a challenge because platinum is so dense. This repeater had the most incredible sound I have ever heard. Just incredible. It was absolutely beautiful, better than anything you could get from a rose-gold case, for instance. I’d never heard anything like it, and it should have been impossible in a platinum case. Naturally we wanted to figure out why it sounded so wonderful, so we sent it back to the lab, we took it apart, put it back together — we analyzed every part of its structure and function.”
“Did you figure out why it sounded so good?” we asked.
He shrugged. “No, we never did. We still have it though. Maybe, someday…”