Beat error is measured in milliseconds, and it describes the symmetry of the balance wheel’s swing. In each oscillation the balance travels one direction (tick) and then the other (tock). In a perfect movement those two half-swings take exactly the same time. Beat error is the difference between them. A reading of 0.0 ms means perfect symmetry. A reading of 2.0 ms means one direction takes two milliseconds longer than the other.
A balance spring next to a human hair, for scale. Its centering is the usual cause of beat error. Photo: Wikimedia Commons, CC BY-SA 4.0.
What it actually measures
The beat error number is a report on geometry. For the tick and tock to be equal, the hairspring must be centered on the balance staff, the impulse jewel must sit correctly between the pallet fork horns, and the whole balance assembly must be poised so neither side of the swing has an advantage. When the hairspring is mounted even slightly off center, or the stud that anchors its outer end sits a fraction out of position, one half-swing gets a head start and the other pays for it.
This is why beat error often shifts after a movement is disassembled and reassembled, even when nothing was damaged. Removing and replacing the balance can move the hairspring stud by an amount invisible to the eye and large to the escapement. Watchmakers correct beat error by rotating the hairspring collet or repositioning the stud, which is delicate work at a scale where a sneeze matters. It is not a user adjustment.
A balance wheel and spring in a 1950s alarm clock, with the regulator pins visible. The spring’s position relative to those pins affects the beat. Photo: Wikimedia Commons, public domain.
How much it matters
The consensus among watchmakers and experienced collectors is simple: under 1.0 ms is acceptable, and under 0.5 ms is ideal. Those are watchmaking rules of thumb, not an ISO or manufacturer standard. Below those thresholds, beat error is essentially invisible on the wrist. A watch with 0.8 ms of beat error and a steady rate will outperform a watch with 0.0 ms and wandering rate every time, because beat error contributes far less to daily accuracy than rate variation does. It is a diagnostic number, not a performance number.
That said, high beat error is worth investigating because of what it indicates. A watch that arrives at 3 or 4 ms has a geometry problem: a hairspring that got bent during a shock, a balance that was dropped in by someone careless, or a movement that has simply never been properly set up. A beat error that drifts over time, especially combined with falling amplitude, points to a hairspring developing a problem, such as coils beginning to stick from oil contamination or magnetism. The number itself does little harm. The cause behind it might.
Keep digging
Beat error is one of three headline timegrapher numbers. Rate is the average speed, amplitude is the health of the swing, and isochronism is how steady the rate stays as the mainspring winds down. For the energy supply behind all of it, see power reserve.
Why it matters
Beat error matters the way a level foundation matters: you do not notice it when it is right, and everything built on top of it works better when it is. Chasing the number to zero is a watchmaker’s craft exercise, not a wearer’s concern. What the wearer should take from it is simpler. A low, stable beat error means the balance assembly is set up the way its designer intended, and a movement set up correctly is a movement that holds its rate, keeps its amplitude, and comes back from a service behaving itself. It is one small symmetry doing quiet work under the dial.
Thanks for reading this.
