The Bimetallic Balance

A balance wheel keeps time by swinging back and forth at a fixed frequency. The problem is that steel — the traditional material for balance springs — softens with heat. A warmer balance spring loses stiffness, so the spring pushes less urgently, and the watch runs slow. Cool the same watch and it runs fast. Left uncorrected, the error across a voyage from England to the tropics could cost a navigator miles.

The fix arrives in the rim itself. A compensation balance cuts the rim into two or more curved arms, each arm a sandwich of brass bonded to steel. Because brass expands more than steel when heated, the laminated arm bends — specifically, it curls inward, carrying the weighted tip closer to the center. Closer to the center means less rotational inertia, which means the balance swings faster, offsetting the sluggishness of the softened spring. Cool the watch and the arms curl outward, raising inertia to compensate for the stiffer spring. The geometry does the arithmetic continuously, without any intervention.

John Arnold and Thomas Earnshaw both arrived at workable versions of this idea in the final decades of the eighteenth century, a period of fierce priority dispute that the Board of Longitude itself was eventually asked to arbitrate. Arnold's early designs experimented with the positioning of the compensation screws — the small weights threaded into the ends of the arms — to tune the degree of correction. Earnshaw's version was simpler to manufacture and became the basis of nearly all subsequent marine chronometer balances. By the early nineteenth century the bimetallic compensation balance was, for practical purposes, the industry standard.

The device is not perfect. Its correction is tuned to work well across a moderate temperature range, but at extremes it over- or under-corrects. This residual error, sometimes called the "middle temperature error," occupied makers for decades. Various schemes were proposed: additional auxiliary arms, two-piece adjustment weights, differently proportioned alloys in the laminate. None of them eliminated the error entirely; they only narrowed it. The difficulty is that the bimetallic arm bends in a curve, while the ideal compensation is a more complex function of temperature.

When Guillaume developed invar — a nickel-iron alloy with near-zero thermal expansion — in the 1890s, the landscape shifted. A balance made partly or entirely of invar moved far less with temperature, and a spring made from similar low-expansion alloys changed far less in stiffness. The need for a mechanically self-correcting rim diminished. But for a century between Harrison's age and the alloy era, the bimetallic balance was the most elegant answer precision watchmaking had found to a problem that the sea makes merciless.