Even Power From an Uneven Spring

Two mechanisms, centuries apart in refinement, attack the same problem differently.

The Problem and Its Two Answers

A coiled mainspring stores energy mechanically, and it does not release that energy evenly. When freshly wound, the spring is tightly coiled and pushes hard; as it runs down, the torque falls away. The going train — and through it, the escapement — receives a varying impulse across the running period, and a varying impulse produces a varying rate.

The oldest solution is the fusee, a conical pulley connected to the spring barrel by a length of gut or fine chain. When the spring is fully wound, the chain pulls from the narrow end of the cone, giving it less mechanical advantage. As the spring weakens, the chain migrates to the wider end, restoring the leverage. The geometry is chosen so that decreasing spring torque multiplied by increasing radius produces a roughly constant output. Fusees appear in sixteenth-century portable clocks and remained standard in English bracket and fusee-pocket clocks well into the nineteenth century — you will still find them in quality carriage clocks and some precision regulators. The device works, but it adds mass, friction and chain-wear, and it reverses the direction of winding relative to the barrel.

The remontoire is a different approach entirely. Rather than equalising the spring's output directly, it interposes a small, secondary reservoir between the mainspring and the escapement. The mainspring rewinds this reservoir — a light auxiliary spring or small weight — at regular short intervals, typically every thirty seconds or every minute. The escapement draws from this secondary store, which is always fully charged at the moment of release and therefore always delivers the same impulse. The primary spring's variation is confined to the rewinding stroke and never reaches the oscillator at all. Harrison used a remontoire in his wooden-framed clocks; it reappears in high-grade tower clocks and in certain precision regulators where the fusee's friction is considered too great a compromise.

Both devices exist because counting the beats is only as reliable as the force arriving at the escapement. Chain and cone address the problem at the barrel; the remontoire addresses it at the escapement's doorstep. Neither is necessary once a quartz crystal takes over, but in any gear train driven by a wound spring, the uneven coil remains the first thing that must be tamed.