British Museum collections, £12.99
Explore / Articles
Accuracy and the stackfreed
The main cause of inaccuracy in early watches was the great variation in power output of the mainspring as it unwound to drive the watch. A coiled spring produces its greatest energy when fully wound; energy output lessens progressively as the spring uncoils. Because the balance wheel has no natural period of swing, its rate of oscillation is determined by its mass, radius and the amount of energy given to it by the escapement. As a result, without some means of evening out the force exerted by the mainspring, the watch would gain drastically at the beginning of its run and lose significantly towards the end. To ensure reasonable accuracy the watchmakers in the Germanic states introduced the stackfreed, a strong spring with a roller at its free end. The roller acts on a snailed cam, geared to the mainspring arbor. This pushes against the mainspring to lessen its force at the beginning of the run; at the end of the run the stackfreed acts with the mainspring to augment its failing power.
Stop-work on the mainspring ensures that only the middle turns of the spring are used during the watch's run: the difference in power output between the fully wound and the unwound watch is kept to a minimum. The stop-work consists of a pinion on the mainspring arbor that meshes with a wheel on which is mounted a cam. The ratio between the two is about 4:1. The wheel has one uncut space so that it can turn only once when the watch is wound and turn back once as the spring unwinds. Thus winding the watch fully winds the mainspring about four turns where as many as twelve turns might otherwise be available. When the watch is unwound the mainspring still has six turns (the stop-work gears preventing it from unwinding any further). The extremes of the mainspring's power output are therefore unused.