A mainspring is a liar. Fully wound, it shoves hard. Nearly spent, it barely pushes at all. That uneven shove travels straight through the gear train to the escapement, so a watch runs a little fast in the morning and a little slow at night, and no amount of careful adjustment fully erases the difference. Every mechanism in this article exists to fix that one problem: getting a steady, unchanging push from a spring that never pushes steadily.
Diagram of Girard-Perregaux’s Constant Escapement, with its silicon buckling blade at the center. Photo: Wikimedia Commons, CC BY-SA 3.0.
The problem
The technical name for the mainspring’s dishonesty is the torque curve, and it has annoyed watchmakers for centuries. When the balance wheel gets a strong push, it swings wide; when the push weakens, the swing shrinks, and a smaller swing takes a slightly different amount of time. That is why power reserve matters for accuracy, not just convenience: a watch is at its most truthful with a full wind and gets progressively more creative as the spring runs down. The whole game of constant force is flattening that curve, so the escapement never knows how much wind is left.
The old answers
The oldest solution is the fusee-and-chain, a cone-shaped pulley that changes the leverage as the spring unwinds: full spring pulls the narrow end, weak spring pulls the wide end, and the delivered force stays even. It worked beautifully for centuries and finally retired in the 1800s, when the going barrel made it unnecessary. The other classic is the remontoir, a tiny secondary spring near the escapement that the mainspring rewinds at fixed intervals, every second or every few seconds, so the escapement only ever feels the little spring’s steady push. And the blunt modern answer is simply more barrel: twin or stacked barrels that share the load and stretch the flatter middle of the torque curve across days instead of hours.
The modern answer
In 2008, Girard-Perregaux showed a concept watch with a stranger idea: build the constant force into the escapement itself. The Constant Escapement uses a silicon blade, six times thinner than a human hair, that buckles and snaps back with exactly the same force every time, feeding the balance a perfectly uniform impulse no matter what the barrels are doing. The idea came from watchmaker Nicolas Déhon, who noticed the way a card snaps between two curved shapes, and it took years of development to make it work reliably. The production version, the Constant Escapement L.M., arrived in 2013 and won the Aiguille d’Or at the Grand Prix d’Horlogerie de Genève that year. A revised Neo Constant Escapement followed in 2023. It is the most radical rethink of the problem since the fusee, and like the fusee, it solves it with geometry instead of brute force.
The Constant Escapement L.M. movement, which Girard-Perregaux put into production in 2013. Photo: Wikimedia Commons, CC BY-SA 3.0.
Why it matters
Constant force sounds like a footnote until you realize the entire history of precision watchmaking is a war against it. The fusee, the remontoir, the going barrel, twin barrels, silicon blades: five centuries of brilliant people attacking the same dishonest spring. Most watches get by without any of it, and honestly, most wrists would never notice. But the pursuit itself is the point. Every time the industry flattens that curve a little more, the mechanical watch gets a little closer to telling the truth, and the truth is the whole job.
Thanks for reading this.
