Thermal compensation is one of the oldest technical challenges in mechanical watchmaking, and to this day there is no single solution valid for all movements. The principle is simple: any material changes size with temperature, and in a precision movement even variations of a few degrees produce measurable rate errors.
The critical point is the balance wheel with its hairspring. When temperature rises, the metal softens slightly and the hairspring loses rigidity: the balance wheel oscillates more slowly, and the watch runs slow. The first historical solution was the bimetallic balance wheel, already known in the eighteenth century: the rim is composed of two metals laminated together — typically brass and steel — with different coefficients of expansion. The rim is cut at two points. With heat, the two halves curve toward the center, shifting mass toward the axis and accelerating the balance wheel just enough to compensate for the relaxation of the hairspring. The Calibre 135 Observatoire featured a split bimetallic balance with adjustable inertia weights: maximum refinement, but also maximum sensitivity to any variation in regulation. Harrison had already applied the same principle to the pendulum of his H1, exploiting materials with different expansion rates to keep the effective length of the pendulum constant even at sea.
The twentieth century shifted the problem from balance wheel geometry to hairspring materials. Nivachron and, later, silicon have nearly eliminated thermal drift of the spring. Patek Philippe's Spiromax, developed with CSEM and introduced in 2006, uses layers of silicon dioxide on the surface of the silicon hairspring to achieve a thermal elasticity coefficient close to zero. The Glucydur balance wheel — a beryllium-copper alloy — works in the same direction: high density, low expansion, and in the free-sprung version it also eliminates the uncertainties introduced by the regulating pin.
In quartz movements the problem transforms: the quartz crystal changes resonance frequency with temperature in a predictable and measurable way. A thermocompensated quartz — such as the ETA G10.212 AD — continuously reads temperature and electronically corrects the frequency, achieving accuracies on the order of ±10 seconds per year that no mechanical balance wheel can approach in daily use.
In fluid systems like the Ressence Type 3 and HYT watches, the problem is physically different: the oil or fluid expands with heat and creates internal pressures that distort the reading or risk detaching the crystal. The solution is compensating bellows — seven in the Type 3 — that absorb the volume variation without transferring pressure to the crystal. In HYT mainsprings the compensating bellows is integrated into the hydraulic system itself. From the invar pendulum to silicon hairsprings, each era has found its own language to answer the same physics.