A mainspring is a coiled spiral torsion spring that stores mechanical energy when wound and releases it gradually to drive a clock, watch, or spring-powered mechanism. The earliest documented use traces to German clockmaker Peter Henlein around 1510, who used coiled steel ribbons to power the first portable clocks. As you wind the spring, you twist it tighter against an arbor inside a barrel; as it unwinds, it delivers torque to a gear train. Modern alloy mainsprings like Nivaflex run 40+ hours per wind and survive millions of cycles without taking a permanent set.
Mainspring Interactive Calculator
Vary mainspring ribbon size, peak elastic strain, and reserve time to see torque, stored energy, winding turns, and average delivered power.
Equation Used
This estimates a flat coiled mainspring as an elastic rectangular strip. Width, thickness, length, and peak strain set the ideal torque and stored bending energy; the selected run time converts that energy into average delivered power.
- Ribbon behaves as a rectangular elastic bending strip with nearly constant moment.
- Elastic modulus is fixed at E = 200 GPa for spring steel style estimation.
- Peak strain remains in the elastic range and no coil binding or set is included.
- Torque is an ideal peak estimate, not a full real-world watch torque curve.
How the Mainspring Actually Works
A mainspring is a long, thin ribbon of spring steel or cobalt-nickel alloy coiled inside a cylindrical housing called the barrel. One end hooks to a central arbor, the other end hooks to the barrel wall. When you wind the watch, the arbor turns and twists the spring inward, storing energy as elastic strain. Release the arbor and the spring tries to uncoil, dragging the barrel wall around with it — that rotation drives the gear train. A typical wristwatch mainspring is 300-400 mm long, 1.0-1.5 mm wide, and only 0.10-0.13 mm thick. The going barrel must hold all that ribbon plus enough free space for it to expand as it unwinds, otherwise you get coil binding and the watch stops with hours of reserve still locked inside.
The physics looks simple but the torque curve is the thorn. A spiral torsion spring delivers more torque when fully wound than when nearly run down — sometimes 2x more. That variation kills timekeeping accuracy in a clock because the escapement loses amplitude as torque drops. Early clockmakers solved this with the fusee chain, a tapered cone that gives the spring more leverage as it weakens. Modern watches use stopwork mechanisms like the Maltese cross to limit the wind to the middle, flatter portion of the torque curve, plus they rely on a balance wheel design that's relatively isochronous across the working amplitude range.
Get the dimensions wrong and the spring fails in predictable ways. A spring too thick for its barrel coil-binds before delivering full power reserve. A spring set too hard during winding develops a permanent kink — what watchmakers call "taking a set" — and you lose 20-30% of usable torque. Modern alloys like Nivaflex NM, made of cobalt, nickel, chromium and beryllium, resist set almost entirely and tolerate the 2-3% strain at peak wind without yielding. The old blue steel mainsprings, by comparison, are charming but break — and when they snap, they snap violently inside the barrel.
Key Components
- Spring Ribbon: The energy-storing element itself, typically 0.10-0.13 mm thick for wristwatches and up to 0.4 mm for marine chronometers. Modern alloys like Nivaflex NM tolerate 2-3% peak strain without taking a set, while older carbon-steel ribbons begin yielding above 1.5%.
- Barrel: The cylindrical housing that contains the spring and acts as the output gear. Diameter ranges from 8-12 mm in wristwatches to 60+ mm in tower clocks. Internal volume must give the unwound spring 50-60% free space, otherwise coils bind and the spring delivers only partial power reserve.
- Arbor: The central shaft the spring hooks onto and around which it gets wound. Surface finish on the arbor matters — a rough arbor scores the inner coil and creates fatigue cracks that propagate over thousands of winds.
- Bridle (slipping spring): Used in automatic watches, this is a curved end-piece that grips the barrel wall by friction. Once peak torque exceeds the bridle's grip force, it slips, preventing overwind. Set the slip threshold too low and the watch loses reserve; too high and rotor inertia stresses the train.
- Stopwork: A geared cam mechanism — classically a Maltese cross — that limits the wind to the middle 4-6 turns of the spring's available range. This trims off the high-torque overwound portion and the low-torque exhausted portion, leaving only the flattest part of the torque curve.
- Fusee and Chain: An equalising mechanism on high-grade clocks and chronometers. The chain wraps around a tapered cone, giving the spring more mechanical advantage as it weakens. A well-tuned fusee holds output torque within 2-3% across the full power reserve.
Real-World Applications of the Mainspring
Mainsprings show up wherever you need stored mechanical energy without electricity — predominantly horology, but also wind-up toys, mechanical music boxes, kitchen timers, and emergency backup mechanisms. The spring you find in a kid's wind-up car runs on the same principle as the one in a Patek Philippe perpetual calendar — only the alloy and tolerances differ.
- Horology — Wristwatches: The Rolex Caliber 3235 uses a Nivaflex NM mainspring delivering 70 hours of power reserve in a single barrel.
- Marine Chronometers: Mercer and Hamilton chronometers used fusee-equalised mainsprings to maintain rate stability of ±0.5 seconds per day across an 8-day reserve.
- Mechanical Music: Reuge cylinder music boxes use flat carbon-steel mainsprings driving a 144-tooth comb at 30-40 RPM playback speed.
- Wind-up Toys: Schylling tin wind-up toys use stamped-steel mainsprings 0.2 mm thick, sized for 8-12 seconds of run time per full wind.
- Mechanical Kitchen Timers: Lux and Robertshaw spring-driven timers use a flat coil mainspring driving a verge escapement, typically for 0-60 minute countdown ranges.
- Tower Clocks: Smaller turret clocks like Smith of Derby installations use multi-barrel mainspring drives where weight-driven systems are impractical.
- Emergency Mechanisms: Mechanical egg timers, vintage camera shutters such as the Compur leaf shutter, and clockwork radios like the Trevor Baylis Freeplay design.
The Formula Behind the Mainspring
The headline number for any mainspring design is power reserve — how many hours the watch runs on one full wind. It comes down to four things: spring length, thickness, barrel diameter, and the gear ratio between barrel and escapement. At the low end of typical operating range — short, thin springs in fashion watches — you get 30-36 hours. At the nominal end you get 40-50 hours, the design sweet spot for a manual wind. At the high end, with extended barrels or twin-barrel designs, you cross into 70-100+ hours, but you pay in case thickness and winding effort. The formula below estimates power reserve from physical dimensions and the train ratio.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Treserve | Power reserve duration on a full wind | seconds | seconds |
| Lspring | Effective working length of the mainspring ribbon | mm | in |
| Darbor | Diameter of the central arbor (the inner winding diameter) | mm | in |
| Rtrain | Gear ratio from barrel to escape wheel | ratio | ratio |
| fescape | Escape wheel beat frequency | Hz | Hz |
Worked Example: Mainspring in a hand-wound dive watch caliber
A small independent watchmaker in Glashütte is sizing the mainspring for a hand-wound dive watch caliber targeting a 48-hour power reserve. The barrel measures 11 mm outer diameter, the arbor is 3.2 mm, the train ratio from barrel to escape wheel is 2880:1, and the escapement runs at 4 Hz (28,800 vph). The spring ribbon is Nivaflex NM at 0.11 mm thickness and 1.20 mm height.
Given
- Lspring = 380 mm
- Darbor = 3.2 mm
- Rtrain = 2880 ratio
- fescape = 4 Hz
- Barrel ID = 9.6 mm
Solution
Step 1 — calculate the number of arbor turns the spring can deliver. The spring length divided by arbor circumference gives total available turns:
In practice you only use the middle portion. Stopwork or just careful design limits the working range to about 6-7 turns of the barrel, which is what reaches the escapement after the train ratio is accounted for.
Step 2 — at nominal design with 6.5 working turns of the barrel, compute escape wheel beats delivered:
So the nominal 48-hour target is hit with 6.5 working barrel turns. That's exactly where the watchmaker wants to land — long enough to survive a weekend off the wrist, short enough to keep the spring in its flat-torque middle range.
Step 3 — at the low end of the range, suppose the spring takes a slight set after 2 years and effective length drops 8% to 350 mm. Working turns fall to about 6.0, and reserve drops:
The owner notices — a Friday-evening wind no longer survives until Sunday morning. At the high end, a fresh spring with the bridle slipping perfectly might deliver 7.0 working turns before the stopwork engages:
That extra 2.4 hours is the comfortable margin. The watch hits 48-hour reserve with headroom, and amplitude stays above 270° across the entire run — which is what keeps timekeeping inside ±4 s/day.
Result
The nominal power reserve is 48 hours with 6. 5 working turns of the barrel. In practice that means the watch will run from Friday evening to Sunday evening on a single wind, with amplitude staying well above the 220° threshold where rate accuracy starts to drift. At the low end after spring fatigue, reserve drops to roughly 43 hours; at the high end with a fresh, perfectly-tuned bridle, you see closer to 50 hours — a useful 7-hour spread that defines the spec window. If your prototype measures only 38 hours instead of 48, check three things in order: (1) coil binding from a barrel with insufficient free space — the spring should occupy 50-55% of barrel volume unwound, not 70%; (2) bridle slip threshold set too low, dumping energy as friction heat instead of storing it; (3) end-curve geometry on the outer hook — a sharp bend instead of the proper S-curve concentrates stress and limits how many turns the spring tolerates before yielding.
Choosing the Mainspring: Pros and Cons
A mainspring is one of several ways to store mechanical energy for a timepiece or wind-up device. The two main alternatives — gravity-driven weights and electric battery drives — each occupy a different niche. Pick wrong and you build something either too bulky to wear or too sensitive to environment.
| Property | Mainspring | Gravity-driven weight | Quartz battery |
|---|---|---|---|
| Energy density | ~0.5-1 J per gram of spring (modern alloys) | Very low — depends on drop height available | ~150 J per gram (CR2032) |
| Run time per charge | 40-100 hours typical, up to 31 days in extended-reserve calibers | 8 days standard for grandfather clocks | 2-5 years per battery |
| Position sensitivity | Insensitive to orientation — works on the wrist or in a pocket | Requires vertical orientation, fixed mounting | Insensitive |
| Torque consistency across discharge | Drops 30-50% from full to empty without stopwork; ±2-3% with fusee | Constant — gravity does not vary | Constant until cutoff |
| Lifespan / fatigue cycles | 10-20 years before set or breakage in steel; 30+ years in Nivaflex | Indefinite — no fatigue | 5-10 years before electronics fail |
| Typical cost (component level) | $5-50 for a watch mainspring; $200+ for matched fusee springs | $10-100 for cast iron weights | $0.50-2 per battery |
| Best application fit | Portable timepieces, wind-up toys, sealed mechanical devices | Stationary clocks where size and weight don't matter | Low-cost or high-accuracy electronic timepieces |
Frequently Asked Questions About Mainspring
Mainspring torque drops as it unwinds, and below a certain torque threshold the balance wheel amplitude falls below 220°. Once amplitude drops, the escapement geometry changes — the impulse pin contact shifts, isochronism breaks down, and the watch starts losing 5-15 seconds per day in the final hours.
This is why high-grade calibers use stopwork to limit operation to the flat middle portion of the torque curve, and why marine chronometers use fusees. If you want consistent rate, wind the watch daily and ignore the last quarter of the reserve — that's the design intent for most non-fusee movements.
Twin barrels in series double the available spring length without doubling barrel diameter, so you keep the case thin. The Lange 31 uses two huge barrels in series for 31 days of reserve. Twin barrels in parallel double the torque, which is useful for driving complications like chronograph chains or tourbillon cages, not for extending reserve.
Rule of thumb: if you need more than 60 hours of reserve in a sub-12 mm thick case, go series twin-barrel. If you need to drive a heavy complication without amplitude collapse, go parallel.
You almost certainly have the wrong strength — height (width) and thickness don't match the original spec. A spring 0.02 mm too thin delivers proportionally less torque and the watch trips through the barrel turns faster because the bridle slips earlier or the stopwork was designed for a stiffer ribbon.
Check the original spring's stamped strength code (most Swiss springs carry a GR or Generale Ressorts code). If you bought a generic replacement by outer diameter alone, you're guessing on three dimensions. Order by exact caliber number — Ronda, ETA, and Sellita publish spring specs by movement.
A spring takes a set when it's stored fully wound for long periods, or wound past its elastic limit. The metal yields slightly and the spring no longer returns to its original free shape. You lose 10-30% of usable torque — the watch winds normally but runs short.
Diagnostic check: pull the spring out of the barrel. A healthy modern spring should spring open into a wide spiral roughly 2-3x the barrel diameter. A set spring stays tightly coiled, looking like it's still inside the barrel. If you see that, replace it — there's no recovering a set spring.
Stronger springs damage movements. Higher torque accelerates wear on every pivot in the gear train, increases impulse force on the escape wheel teeth, and pushes balance amplitude past 320° where rebanking starts — the impulse pin slams into the wrong side of the fork horns and the watch can stop or run wildly.
Poor amplitude is almost always a lubrication or wear issue, not a torque issue. Service the movement first. Only step up spring strength if you've genuinely got a movement designed for a stronger spring (some calibers ship with multiple strength options for different complications).
That's a fusee profile mismatch. The cone is cut to match a specific spring's torque curve — if the spring has been replaced with a slightly different strength, or if the chain has stretched (they do, by 1-2% over decades), the leverage at one specific point in the run no longer compensates for the spring's torque, and the train doesn't get enough drive to overcome friction at that exact rotation.
Mark the fusee at the stopping point and check chain wear at that spot first. Replacement chains from clockmakers like Cousins UK come in matched lengths, but the original fusee profile assumes the original chain — old clocks rarely accept modern chains without recutting the fusee.
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.
