When you flip the switch and electricity flows through the wire to the light bulbs in your home, you are hardly aware of the fact that at the other end of the wire, there’s an electrical generator putting out hundreds of thousands of volts of power, which has to be converted into a controlled, steady flow of energy before it can be of any use. Likewise in a watch, the hypnotic and steady back-and-forth swinging of the balance is the end result of the power locked up in the mainspring, coiled like a tiger ready to spring, but hidden from view inside the barrel. The potentially explosive energy of the mainspring has to be controlled for it to be useful, otherwise the result would be the same as if you tried to plug your toaster directly into a power station — more than bread would go up in smoke.
The heart of the problem is that for a watch to do what most of us expect it to do — keep reasonably good time — the size of the oscillations of the balance wheel have to fall within a pretty narrow range. Too much power to the balance will simply send it swinging wildly around in a full circle, knocking the delicate impulse jewel against the pallet fork; too little, and the balance will listlessly wobble to-and-fro like a sullen child on a playground swing. Both will have disastrous consequences for accuracy, but the latter is the more insidious problem as the lack of energy makes the watch much more susceptible to variations in rate, depending on the position the watch is held in. The wrist, it should be borne in mind, is actually a pretty ridiculous place to keep a watch, so much so that as late as the early 1930s, German watchmakers derided wristwatches as a temporary and preposterous fad, and hoped that watch owners — especially male watch owners, who, one hoped, could be counted on to be sensible and not be taken in by the fashion for an effeminate affectation — would soon return to the custom of wearing good watches in a waistcoat pocket, where they belonged. “The idiotic fashion of carrying one’s clock on the most restless part of the body… will, one hopes, soon disappear,” opined one Professor Bock of Hamburg.
The good professor’s pedantic irritation might seem laughable today but it’s rooted in the inescapable fact that as a watch assumes different positions, the effect of gravity on the escapement causes the watch to speed up or slow down — for instance, if your watch is held perpendicular to the ground, the tiny steel pivots of the balance wheel rest their entire length in the jeweled bearings they run in. This miniscule increase in friction actually suffices to make the watch run a little fast by decreasing the time of each oscillation. Truly high-quality watches have always been judged by how little variation there is in rate as the watch changes position. In fact, the Contrôle Officiel Suisse des Chronomètres (COSC) — the official certifier of chronometers — in part defines a chronometer in terms of its ability to run at close to the same rate, no matter what position the watch is in.
What this means to an intrepid horological engineer attempting to design a watch with a long power reserve is that unless some way can be found to even out the flow of power from the unwinding mainspring to the balance, the watch, as it runs through its power reserve, will increasingly tend to run at different rates in different positions. An extended power reserve in such a case simply means a longer and longer period in which to see your watch keep worse and worse time, which rather takes the fun out of it. For a watch that starts out running accurately at the beginning of its power reserve to do so at the end, the tendency of a mainspring to deliver excess power initially and inadequate power as it approaches the end of its reserve has to somehow be modulated.

Building a Better Beast: Modern Mainsprings In a perfect world, variations in mainspring power shouldn’t make any difference in rate, but as anyone who has ever sent a watch to a brand service center knows, we don’t live in a perfect world. The combination of a balance and a spring should have a property known as ‘isochronism’ — literally, “equal time” — that is, it should take the same amount of time for the balance to go back and forth, no matter how big or small the oscillation is. Pendulums have a similar property — the length of time it takes for a pendulum to swing is not a function of how far it swings, but rather how long the pendulum is. Our sullen child swinging a listless few degrees to-and-fro takes the same amount of time to do so as the irrepressible moppet in the next swing to go through an alarming 180 degrees of arc.
In the earliest watches, mainsprings were made of ordinary carbon steel, and such mainsprings could be relied on to deliver wildly variable power as they unwound, and also to lose their elasticity quite rapidly; so much so, in fact, that many early watches actually have a worm-gear device accessible without opening the watch that was used to pick up the slack in the mainspring, thus saving the watchmaker the trouble of swapping out the old mainspring too frequently. Modern watches benefit from mainsprings made of alloys that are far less prone to losing their elasticity over repeated windings and unwindings (to say nothing of being unlikely to actually break — an all too frequent occurrence in the not-so-good old days).
In addition, modern mainsprings are shaped differently: when out of the barrel, a carbon steel mainspring is a relaxed spiral; but a modern mainspring has an outer curve that is the reverse of the inner curve — out of the barrel, it forms an ‘s’ shape, rather reminiscent of a G clef. The effect of this reverse curve is to hold the inner coils of a fully wound spring closer to the center of the barrel, away from the outermost coil of the spring, reducing the friction between the coils created as the spring unwinds. This also helps the spring deliver power more evenly.
For a modern watch, with an alloy mainspring that is wound every day, the use of only the first 24 hours (more or less) of the reserve, plus the inherently greater resistance to positional error afforded by modern escapements and methods of construction, generally suffices to produce at least the potential of satisfactory accuracy — potential, because carelessness in manufacturing can easily produce a watch with unnecessarily poor performance. For modern automatic watches worn every day, the issue of variability in mainspring power is even less germane as the presence of a slipping bridle on the outer coil of the mainspring prevents excessive tightness, helping to reduce exposure of the escapement to excessive power, while the automatic winding mechanism prevents use of the last and weakest coils of the mainspring, ensuring a generally fairly even power flow to the escapement, without requiring recourse to any other, more exotic measures.

Monster on a Leash: Taming a One-Month Watch For an ordinary watch that will run an ordinary number of hours — 36 to 40 on average — the use of modern escapements and alloys suffices, but for the new generation of long power reserve watches, such expedients are inadequate. The colossal power of the mainsprings used in long power reserve watches, some of which take up the entire diameter of the case, can exhibit enormous fluctuations between the first and last days of power delivery, and in order for a watch that runs long to run well, steps must be taken to tame the unruly beast in the barrel.
One strategy to adopt is to simply put in stopworks that keep the last day of power reserve from being used — you don’t have to worry about low energy from the last eighth of the power reserve if the watch only runs for seven days! Couple this with an automatic winding mechanism and you have an excellent chance of producing a watch that will keep a good rate, even if left running but unworn for several days.
For example, the combination of stopworks plus an automatic winding system is found in the IWC cal. 50010 — the watch has a power reserve indicator showing seven days, but there are actually eight days’ worth of juice in the tank.
Another useful design approach is to use multiple barrels. In the same way that a team of horses is easier to control than a single wild bull, multiple barrels moderate each other’s power delivery. In the new Blancpain cal. 13R0 for example, three mainspring barrels achieve eight days of power reserve. The three mainspring barrels run in series, with one actually delivering power to the gear train and the other two storing additional energy, which is transferred to the primary mainspring barrel via intermediate gears. This allows the watch to maintain a more consistent degree of energy delivery to the escapement over the duration of its running time.
At the extreme limit of long power reserves, however, the potential energy necessary to keep a watch running for weeks rather than days calls forth, in an act of horological necromancy, solutions which are so rare that they are almost unheard of in modern wristwatches. Such a problem confronted the designers of the Lange 31.
Maintaining an even delivery of power over a 31-day period from two huge mainsprings seems an impossible task; the use of multiple barrels would be unsatisfactory as the result would have been something so large that it can be a wristwatch in name only.
The first possibility to present itself under such circumstances is the chain and fusée. The chain and fusée, in which a chain unwinds from a spiral grooved cone onto the mainspring barrel, is an excellent solution to the problem of maintaining steady power flow to the escapement; so much so that in English watchmaking, fusées have been nearly ubiquitous in high-grade watches for many decades. Found wherever no effort is spared to produce as stable a rate as possible, from regulators to marine chronometers, the fusée would have been a perfect choice but for one problem — a lack of space.
With the two mainsprings housed in barrels already 25 mm in diameter, a chain and fusée arrangement would not only have produced a watch of unsatisfactory width, it would have made for an exaggerated height as well, and when you’re trying to wring a month’s worth of ticks and tocks out of a watch, anything that makes the movement beefier than it is already, is not a good thing.
Another possibility might have been to resurrect a device known as the stackfreed, used in some early German watches. A stackfreed consists of a cam attached to the mainspring barrel, with a tension spring applying pressure to it. The cam is shaped such that the greater pressure is applied during the first part of the watch’s power reserve, the pressure lessening as the spring weakens. While the stackfreed works, it is an unsatisfyingly crude solution — much like braking a car by chaining a ship’s anchor to the rear bumper and throwing it out the window at every stoplight.
Instead, the designers at A. Lange & Söhne chose another, and even rarer complication: the remontoir.
The remontoir (or remontoir d’égalité, to give it its full name) is a device designed to ‘step down’ the energy of the mainspring, somewhat like how a transformer steps down the voltage of current coming from an electrical main before it reaches the consumer. To do this, a spring — usually either blade-shaped, flat, or a spiral spring similar to a hairspring — is placed somewhere in the gear train between the mainspring barrel and the escapement. The escapement, therefore, is not directly driven by the mainspring; instead, the mainspring simply serves to keep the remontoir spring ‘armed’ or wound up, so that it always delivers exactly the same amount of energy to the escapement, regardless of the output of the mainspring.
The remontoir, like the tourbillon, was originally intended as an enhancement to chronometric performance and, like the tourbillon, is generally thought to be difficult and demanding to construct. Remontoirs are also, in general, not needed for a watch to exhibit excellent performance. Although, in the words of George Daniels, “The fact that the mechanism (a remontoir) is unnecessary merely adds to its charm.” However, in a watch designed to run for a month, the remontoir comes into its own.
For the Lange 31, a unique design was developed, in which a three-lobed cam rotates between the jeweled jaws of a lever very reminiscent of the pallet fork of the Swiss lever escapement. As it rotates, this lever moves slowly back and forth, and as it does so, it alternately holds and releases a single-toothed wheel driven by the power train. When unlocked by the remontoir lever and allowed to rotate, this single-toothed wheel winds up the remontoir spring — a spiral spring, in this case — which directly drives the fourth wheel. The fourth wheel then drives the escape wheel in the usual fashion.
Like the tourbillon, the fusée and remontoir are in general, on the very rare occasions that they are seen, manifestations of skill rather than necessity (not a particularly bad thing, since the same thing could arguably be said about mechanical watches in general). But in a watch with an extremely long power reserve, the need for such archaic yet fascinating complications reasserts itself, and like a stallion broken to the saddle, the tremendous energy necessary to drive these horological marathoners can be tamed. H

For a watch to keep reasonably good time, the oscillations of the balance wheel have to fall within a pretty narrow range

IWC caliber 5000

The Lange 31 uses a remontoir

The chain and fusée, in which a chain unwinds from a spiral grooved cone onto the mainspring barrel, is an excellent solution to the problem of maintaining steady power flow to the escapement

the combination of stopworks plus an automatic winding system is found in the IWC Cal. 50010 — the watch has a power reserve indicator showing seven days, but there are actually eight days’ worth of juice in the tank

Breguet La Tradition tourbillon with chain and fusée

F.P. Journe Tourbillon Souverain

The double barrel caliber used in Omega’s Hour Vision

Also Read