A barrel in horology is the cylindrical container inside a mechanical watch or clock that houses the coiled mainspring. The mainspring itself is the critical component — a hardened steel ribbon that stores energy when wound and releases it gradually as it uncoils, driving the barrel teeth which mesh with the first wheel of the going train. The barrel exists to convert a single winding event into hours of regulated torque output. A modern wristwatch barrel typically delivers 38 to 80 hours of power reserve from one full wind.
Barrel (Horology) Interactive Calculator
Vary barrel size, arbor size, spring length, and spring thickness to see the mainspring fill ratio and whether it lands near the 50 to 55% watch-barrel target.
Equation Used
The calculator compares the top-view area of the coiled mainspring strip, L times t, with the usable annular area inside the barrel after the arbor is removed. A fill ratio near 50 to 55% matches the article guidance: too little spring reduces reserve, while too much can prevent clean uncoiling.
- Mainspring height is close to barrel internal height, so height cancels from the volume ratio.
- Spring strip is modeled as a rectangular strip with area L * t in top view.
- The arbor diameter is excluded from the usable barrel annulus.
- The preferred fill target is 50 to 55% as stated in the article.
Inside the Barrel (horology)
The barrel is a low, thin-walled drum — usually 10 to 14 mm in diameter on a wristwatch — closed by a press-fit lid and rotating on a central arbor. Inside, the mainspring is coiled around the arbor with its inner end hooked to the arbor itself and its outer end hooked or slipping against the barrel wall. Wind the watch and the arbor turns while the barrel stays still; this winds the spring tighter against the arbor. Once you let go, the click and ratchet wheel hold the arbor fixed, and now the spring pushes against the barrel wall, rotating the barrel slowly. The barrel teeth, cut into the outer rim, drive the centre wheel pinion and start the going train turning.
Geometry matters more than people expect. The barrel must be filled to roughly 50 to 55% of its internal volume by the mainspring — too little and the power reserve drops below spec, too much and the spring cannot uncoil cleanly and torque collapses early. The mainspring thickness on a typical 11½ ligne calibre runs 0.10 to 0.13 mm, with a height equal to the inside of the barrel minus about 0.02 mm clearance. Get that clearance wrong and you see two failure modes: a tight spring scrapes the lid and produces erratic amplitude, while excess play lets the coils slope and bind, which shows up on a timegrapher as wandering rate over the first 12 hours.
The most common real-world failures are a slipping bridle on automatic-winding barrels (caused by a glazed barrel wall or worn bridle, leading to under-winding even with full rotor activity), a broken arbor hook (which strips the inner coil and kills power instantly), and lid deformation from over-pressing during service. If amplitude drops below 200° at full wind on a healthy escapement, suspect the barrel before the hairspring.
Key Components
- Barrel drum and lid: The cylindrical housing that contains the mainspring and carries the gear teeth on its outer rim. Wall thickness is typically 0.30 to 0.40 mm on a wristwatch barrel; the lid press-fits with about 0.02 to 0.03 mm interference. Concentricity of the lid bore to the drum bore must hold within 0.01 mm or the barrel wobbles and load fluctuates.
- Mainspring: A pre-formed strip of hardened, blued or alloy steel (Nivaflex is the modern standard) coiled around the arbor. Length runs 250 to 450 mm in a wristwatch, thickness 0.09 to 0.13 mm. Its S-curve pre-set determines the torque curve — a flat-set spring delivers high initial torque that drops fast, while a reverse-curve spring trades peak torque for a flatter delivery.
- Barrel arbor: The central shaft the mainspring's inner coil hooks onto. The arbor hook is the single highest-stress feature in the entire movement. The arbor pivot diameters are typically 0.40 to 0.60 mm and run in jewelled bearings in the mainplate and barrel bridge.
- Bridle (slipping spring): Fitted on automatic-winding watches only. The outer end of the mainspring carries a curved bridle that grips the barrel wall under normal torque but slips when the spring is fully wound, preventing over-tensioning. Slip torque is set at roughly 1.3 to 1.5× the working torque.
- Ratchet wheel and click: Mounted on the squared end of the arbor on top of the barrel bridge. The click engages the ratchet wheel teeth to hold the arbor against the wound spring. Click-spring force runs 0.05 to 0.15 N — enough to drop the click cleanly between teeth without dragging.
- Barrel teeth: Cut into the outer rim of the barrel drum, typically 60 to 100 teeth on a wristwatch barrel. Tooth profile is normally a horological cycloidal form. The teeth mesh with the centre wheel pinion (or first wheel of the going train) to deliver torque to the movement.
Where the Barrel (horology) Is Used
Every spring-driven mechanical timepiece uses a barrel of some form, but the specific design varies widely with the application. Long-running clocks use larger barrels with longer mainsprings, marine chronometers use fusee-coupled barrels for torque uniformity, and high-beat wristwatches often use twin or triple barrel arrangements to combine extended power reserve with stable amplitude. The arrangement of the barrel — going barrel, hanging barrel, or fusee-coupled — directly sets how the watch behaves over its power-reserve window.
- Wristwatch manufacturing: ETA 2824-2 movement uses a single going barrel delivering 38 hours of reserve at 28,800 vph; the workhorse calibre in countless Swiss watches from Hamilton to Tudor.
- Haute horlogerie: A. Lange & Söhne Lange 31 uses twin large-diameter barrels with a remontoir to deliver 31 days of constant-force power reserve.
- Marine chronometers: Mercer, Hamilton Model 21 and Ulysse Nardin chronometers historically used a fusee chain coupled to the barrel to flatten the mainspring torque curve to within 1% over 56 hours.
- Tourbillon wristwatches: Greubel Forsey Quadruple Tourbillon uses four serially-coupled barrels providing 72 hours of reserve while supporting four independent tourbillon cages.
- Long-duration clocks: The Atmos clock by Jaeger-LeCoultre uses a barrel wound by an aneroid bellows reacting to ambient temperature changes — a 1°C swing winds enough for 48 hours of running.
- Pocket watches and railroad-grade movements: Hamilton 992B railroad-grade pocket watch uses a going barrel sized for 60 hours of reserve, designed to hold ±3 seconds/day across the full discharge curve.
The Formula Behind the Barrel (horology)
The single number that matters most when sizing a barrel is the power reserve — how many hours the watch runs from a full wind. Power reserve depends on the number of turns the mainspring can deliver, the gear ratio between the barrel and the escape wheel, and the beat rate of the balance. At the low end of typical wristwatch design (around 38 hours, like the ETA 2824-2) you have a thin, short spring and a single barrel — compact but with a steeper torque curve, so amplitude drops noticeably in the last 8 hours. At the high end (120+ hours, like an IWC 8-day calibre) you need either a much longer spring, a larger-diameter barrel, or multiple barrels in series. The sweet spot for a daily-wear automatic sits around 60 to 70 hours — enough to leave the watch off Friday night and put it on Monday morning still running.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| PR | Power reserve, total running time from full wind to stop | hours | hours |
| Nturns | Number of useful turns the barrel makes as the mainspring uncoils (typically 6 to 8) | turns | turns |
| Rtrain | Gear ratio from barrel to escape wheel (escape wheel revs per barrel rev) | dimensionless | dimensionless |
| f | Balance frequency (vibrations per hour, vph) | vph | vph |
| 7200 | Constant: 2 vibrations per escape-wheel tooth × 3600 sec/hour, assuming 15 escape-wheel teeth and standard Swiss lever geometry — adjust if escape wheel tooth count differs | constant | constant |
Worked Example: Barrel (horology) in a 28,800 vph automatic wristwatch calibre
Your team is finalising a new in-house automatic calibre running at 28,800 vph and you need to verify the single-barrel sizing will hit the 60-hour power reserve target. The barrel arbor allows 7.5 useful turns of the mainspring before the bridle starts slipping, and the going train is configured with a barrel-to-escape-wheel ratio of 3375:1 with a 15-tooth escape wheel.
Given
- Nturns = 7.5 turns
- Rtrain = 3375 dimensionless
- f = 28800 vph
Solution
Step 1 — at the nominal target frequency of 28,800 vph (4 Hz, the modern Swiss standard), compute the power reserve directly:
That clears the 60-hour design target with a small margin — exactly where you want it for production tolerance on spring length and slip torque.
Step 2 — check what happens at the low end of the typical operating range. If the bridle wears or the barrel wall glazes and Nturns drops to 6.5 (a common service-interval failure), the reserve falls to:
That's the kind of drop a customer will notice — the watch dies overnight on a weekend instead of carrying through Monday morning. It's the single most common warranty complaint on automatics with marginal barrel sizing.
Step 3 — at the high end, suppose you switch to a 21,600 vph (3 Hz) layout like a vintage-style calibre while keeping the same barrel and train:
Same barrel, same spring, just a slower beat — and you've gained 21 hours of reserve. This is precisely why low-beat watches like the Grand Seiko 9S64 hit 72 hours from a single modest barrel while a high-beat 36,000 vph movement on the same barrel would only manage 50 hours.
Result
The nominal design delivers 63. 3 hours of power reserve, comfortably above the 60-hour target. In the field that means the watch will run from Friday evening through Monday morning with margin — the daily-wear sweet spot. Across the operating range, expect 54.8 hours when the bridle is worn near service interval and 84.4 hours if the same barrel runs at 21,600 vph instead of 28,800 vph, which gives you a feel for how sensitive reserve is to beat rate. If a finished prototype measures 50 hours instead of the predicted 63, check three things in order: (1) mainspring length — a spring cut 8% short during manufacturing knocks roughly 5 hours off reserve, (2) excessive end-shake on the barrel arbor causing the lid to rub the spring's top edge, which dissipates energy as heat across the discharge, and (3) gear-train pivot friction from a dry centre-wheel jewel, which raises torque demand and shortens reserve without showing on a static torque test.
When to Use a Barrel (horology) and When Not To
The barrel is not the only way to feed energy into a mechanical timepiece, and even within barrel-driven movements you have several arrangements to choose from. The decision comes down to torque uniformity, power reserve, complexity, and how much movement real-estate you can afford to give up.
| Property | Going barrel (single) | Twin barrels in series | Fusee-and-chain |
|---|---|---|---|
| Typical power reserve | 38–80 hours | 72–240 hours | 30–56 hours |
| Torque uniformity over reserve | Drops ~30% from full to empty | Drops ~25% (slightly flatter) | Flat to within 1–3% (best in class) |
| Movement real-estate cost | Low — single drum 10–14 mm | Medium — two drums plus coupling | High — fusee cone plus chain plus barrel |
| Manufacturing complexity | Low — standard production | Medium — extra wheel and bridge | Very high — chain has 200+ links, hand-fitted |
| Service interval impact | Standard 4–5 year service | Standard 4–5 year service | Chain inspection every service, replace at 10 years |
| Cost premium over baseline | Baseline | 1.5–2× baseline | 10–50× baseline (haute horlogerie only) |
| Best application fit | Daily-wear automatics, quartz-equivalent reliability | Long-reserve automatics, perpetual calendars | Marine chronometers, ultra-precision pocket watches |
Frequently Asked Questions About Barrel (horology)
That's the natural shape of the mainspring torque curve, not a fault. A typical going barrel delivers roughly 30% less torque at the bottom of its reserve than at full wind, and the escapement amplitude tracks that torque almost linearly. You'll see 280° at full wind drop to 200° in the final hours.
If the drop is steeper than that — say, amplitude collapsing below 180° with 10 hours nominally remaining — the barrel itself is undersized for the train's torque demand, or the spring's S-curve preset has relaxed. A relaxed spring shows up as low torque across the entire discharge, not just at the end. Replace the mainspring before suspecting the escapement.
Twin barrels in series almost always win at the 100-hour target. A single barrel sized for 100 hours needs either an unusually long spring (which forces a larger drum diameter and eats movement real-estate) or a thinner spring (which drops peak torque and pushes amplitude down). Twin barrels split the angular travel between two springs, so each spring runs in its more linear range and the combined torque curve is flatter.
The penalty is one extra wheel, one extra bridge, and roughly 1.5× the barrel-region cost. For anything above 120 hours, you're forced into twin or triple barrels regardless — a single barrel can't physically deliver that reserve in a wristwatch footprint.
The bridle is slipping early. On automatic barrels the outer end of the mainspring isn't hooked to the wall — it carries a curved slipping spring (the bridle) that's supposed to grip the wall under working torque and slip only at full wind. If the barrel wall is glazed from old, dried-out braking grease, the bridle starts slipping before the spring reaches its full wound state.
Diagnostic check: wind the watch fully by hand from a stop and time the reserve. If hand-winding gives you the full 60 hours but rotor-winding only gets you to 52, the bridle isn't reaching the same wound position because the rotor's lower peak torque can't overcome the bridle's slip threshold once the spring tightens. Re-grease the barrel wall with fresh Kluber P125 or equivalent braking grease.
More than most people realise. The barrel teeth set the first ratio in the going train, so the count cascades through every wheel after it. A 96-tooth barrel driving a 12-leaf centre pinion gives an 8:1 first reduction; drop to 80 teeth and you need to recover the lost ratio further down the train, usually with a smaller pinion that has its own tooth-strength and depthing problems.
The practical limit is roughly 60 teeth at the low end (below that the tooth load gets aggressive and you see early wear on the centre pinion) and 100 at the high end (above that the teeth get so fine that depthing tolerance becomes critical to within 0.01 mm). Most modern wristwatch barrels sit at 75 to 90 teeth as the sweet spot.
That signature points to the outer coils of the mainspring binding against each other or against the barrel wall as they release stored energy. When the spring is fully wound, the outer coils are pressed hard against the wall; as it unwinds, those coils peel away, and if there's contamination or dried grease between them they release in unpredictable jumps rather than smoothly. Each jump shows on the timegrapher as a brief rate spike.
This is almost always a lubrication issue — either the original braking grease has degraded or the spring was installed without enough fresh grease on the outer coils. After 8 hours the binding coils have separated and the rate stabilises. Strip and re-grease the barrel; don't waste time regulating the hairspring against this signature.
The 50% rule comes from the geometry of an Archimedean spiral. When a fully wound mainspring fills more than about 55% of the available space between the arbor and the wall, there's no room left for the coils to redistribute as the spring releases. The result is that the inner coils stay pinned against the arbor and the spring delivers torque from only its outer turns — you get high initial torque that collapses early instead of a smooth discharge.
In practice, packing in a 10% longer spring than spec doesn't give you 10% more reserve — it gives you maybe 3% more reserve and a much steeper torque curve, which destroys amplitude stability. The 50–55% fill target isn't a guideline, it's a hard physical constraint set by the spiral geometry.
References & Further Reading
- Wikipedia contributors. Mainspring. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
