A geared watch stop is a small lever or yoke that engages the gear train of a watch movement — usually the fourth wheel or the balance rim — to halt motion at a precise rotational position. It works by inserting a thin spring-tensioned blade between two named gear teeth, locking the train against the mainspring's residual torque. Watchmakers use it to synchronise the seconds hand during time setting, freeze the train for rate measurement on a timing machine, or lock the movement during disassembly. The result is sub-tooth positional repeatability, typically within 1 tooth of a 60-tooth fourth wheel — about 6° of arc.
Geared Watch Stop Interactive Calculator
Vary the wheel tooth count and lock span to see the stop resolution, time equivalent, and one-tooth jump risk.
Equation Used
The angular stop resolution is the angular pitch of the target wheel multiplied by the number of tooth pitches accepted as the lock span. For the worked example fourth wheel, N = 60 and k = 1, so theta = 360/60 = 6 degrees per tooth. The time equivalent assumes the fourth wheel carries the seconds hand and turns once per minute.
- Fourth wheel is the stopped wheel and turns at 1 rev/min for the time output.
- Stop repeatability is modeled as an integer number of tooth pitches.
- Blade deflection, tooth backlash, and yoke pivot error are not included.
Inside the Geared Watch Stop
The geared watch stop works on a simple principle: when you pull the crown to the setting position, a stem-driven yoke pushes a slender steel lever into the path of a moving gear. The lever's tip — usually 0.15 to 0.25 mm thick — slides between two adjacent teeth of the fourth wheel, or alternatively rests against the balance rim in a stop-seconds design. The mainspring keeps applying torque, but the train cannot rotate because the lever has converted rolling contact into a static tooth-on-blade interface. Release the crown, the yoke retracts, and the train picks up exactly where it stopped.
The geometry is fussier than it looks. If the blade is too thick it cannot enter the tooth gap and you get a noisy click followed by a stalled lever. If it is too thin it deflects under mainspring torque and the wheel skips a tooth — you set the seconds hand to 12, release the crown, and the hand jumps 6° before it begins running. The typical tolerance on blade thickness is ±0.02 mm relative to the tooth gap. Surface finish matters too: a blade with Ra above 0.4 µm chatters against the tooth flank and leaves marks that show up under 10× inspection.
Failure modes are predictable. The most common is a worn yoke pivot that lets the blade enter at an angle — the tip rides up the tooth face instead of seating in the gap, and the train creeps. Second most common is a fatigued return spring that holds the blade in light contact with the wheel during normal running, robbing the train of amplitude. If you see balance amplitude drop 20° between setting and running, the stop lever is dragging.
Key Components
- Stop Lever (Hacking Lever): A flat hardened steel blade, typically 0.15-0.25 mm thick at the tip, that physically enters the gear train to halt motion. The tip profile must match the tooth gap of the target wheel — for a 60-tooth fourth wheel with module 0.07, the gap width is around 0.18 mm at the pitch line.
- Yoke: The intermediate lever connecting stem position to the stop lever. Pulling the crown to position 2 or 3 rotates the yoke around its pivot, which pushes the stop lever into engagement. Yoke pivot clearance must stay below 0.03 mm or the stop lever enters the train at an angle.
- Return Spring: A flat spring, usually 0.08-0.12 mm thick steel, that pulls the stop lever clear of the train when the crown returns to position 1. Spring force must be high enough to fully retract the blade but low enough that crown pull-out feel stays acceptable — 30 to 60 grams-force at the crown is the typical target.
- Target Wheel: The gear the stop lever engages. In most modern movements this is the fourth wheel because it carries the seconds hand directly — 1 rev/min means stopping it stops visible time. Older designs sometimes stop the third wheel for less precise hacking.
- Setting Stem: Translates crown axial pull into yoke rotation. The detents at positions 1, 2, and 3 must be crisp — a sloppy detent lets the user partially engage the stop lever, which causes the noisy stalled-lever symptom mentioned earlier.
Where the Geared Watch Stop Is Used
Geared watch stops appear anywhere a movement needs to be paused at a known position. The applications split between user-facing functions like seconds hacking on a wristwatch, and workshop functions like immobilising the train during rate measurement or service. The reason the same mechanism serves both roles is that both demand the same thing: a repeatable, non-destructive way to halt a torqued gear train without disturbing escapement geometry. When the stop lever fails, you see two named symptoms — seconds-hand creep on release, or amplitude loss during running — and both trace back to lever geometry or yoke wear.
- Wristwatch manufacturing: ETA 2824-2 movement uses a hacking lever bearing on the balance rim — the standard stop-seconds system found in Swiss three-hand calibres from Tissot, Hamilton, and Mido.
- Marine chronometry: Ulysse Nardin and Hamilton Model 22 deck watches incorporated stop-seconds for synchronising a fleet of timekeepers to a master signal during naval observatory comparison runs.
- Watch service workshops: Witschi Chronoscope X1 timing machines pair with a manually operated stop lever fixture to freeze the fourth wheel between rate measurements without interrupting mainspring torque.
- Military issue watches: MIL-W-46374 specification field watches and the earlier A-11 hacking watch required a functioning stop-seconds for unit time synchronisation before coordinated operations.
- Horological education: WOSTEP and BHI training programs use Unitas 6497 movements as a teaching platform — its visible hacking lever lets students see the stop-seconds engagement geometry directly without removing the dial.
- Movement assembly tooling: Bergeon 5700-Z movement holders include an auxiliary stop-train pin used during disassembly of high-grade calibres like Patek Philippe 215 to safely release mainspring torque tooth by tooth.
The Formula Behind the Geared Watch Stop
The most useful calculation for a geared watch stop is the angular positioning resolution — how precisely you can stop the seconds hand by halting the target wheel. At the low end of the typical range, a stop lever engaging a 30-tooth wheel gives you a coarse 12° resolution — fine for a marine chronometer set to the nearest second but useless for sub-second synchronisation. At the high end, engaging a 90-tooth wheel through a 1:15 reduction gives sub-degree resolution. The sweet spot for most three-hand watches sits at a 60-tooth fourth wheel directly carrying the seconds hand, which gives 6° per tooth — equivalent to 1 second of seconds-hand position.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| θres | Angular positioning resolution at the target wheel — the smallest angle by which you can shift the stopped position | degrees | degrees |
| Nteeth | Number of teeth on the wheel the stop lever engages | teeth (dimensionless) | teeth (dimensionless) |
| tres | Equivalent time resolution at the seconds hand, when target wheel rotates once per minute | seconds | seconds |
Worked Example: Geared Watch Stop in a vintage railway-grade pocket watch service
A specialist horologist in Toronto is overhauling a 1952 Ball Trainmaster 21-jewel pocket watch with a hacking modification added in the 1960s. The fourth wheel carries 60 teeth, runs at 1 rev/min, and the customer wants the seconds-hand sync to land within ±1 second of a GPS time pulse during locomotive engineer certification testing. You need to know what positioning resolution the stop lever delivers and whether it meets the spec.
Given
- Nteeth = 60 teeth
- Fourth wheel rotation rate = 1 rev/min
- Required time accuracy = ±1 second
- Stop lever blade thickness = 0.18 mm
Solution
Step 1 — calculate angular resolution at the nominal 60-tooth fourth wheel:
Step 2 — convert that to seconds-hand position. The fourth wheel turns once per 60 seconds, so 6° equals 1 second of seconds-hand travel:
Step 3 — examine the low end of the typical range. If this were an older movement with a 30-tooth fourth wheel (some 1920s pocket watches used coarser tooth counts):
That fails the ±1 second spec immediately. The hand could land 2 seconds off the GPS pulse with no way to nudge it closer without re-engaging the stop. Step 4 — high end of the typical range, a high-beat calibre like the Zenith El Primero with a 90-tooth fourth wheel:
That comfortably beats the spec but you pay for it in tooth-gap width — the gap on a 90-tooth wheel of similar diameter is around 0.12 mm, demanding a thinner, more delicate stop lever blade that fatigues faster.
Result
The Ball Trainmaster's 60-tooth fourth wheel gives 1. 0 second of positioning resolution per tooth — exactly meeting the ±1 second customer spec, with no margin to spare. In practice this means the user feels a single discrete 'click' position per second when releasing the crown against the GPS pulse — fine enough that a trained operator can hit the pulse on the first try most of the time. The 30-tooth low-end case would force two-attempt synchronisation as a matter of routine, while the 90-tooth high-end case lets you nail it first try every time. If you measure worse than 1 second of repeatability on this movement, check three things in order: (1) a bent stop-lever tip riding the tooth crown instead of seating in the gap — visible under 10× as a polished facet on the lever tip, (2) end-shake on the fourth wheel pivot above 0.04 mm letting the wheel rock axially as the lever lands, or (3) excessive mainspring residual torque on a freshly wound watch causing the wheel to advance one extra tooth after the lever lifts.
Choosing the Geared Watch Stop: Pros and Cons
Three approaches dominate watch stopping: a geared stop lever engaging the train, a friction stop bearing on the balance rim, and a magnetic or electronic hold used in a few quartz and tuning-fork calibres. Each has measurable trade-offs in resolution, drag, and impact on amplitude.
| Property | Geared Watch Stop (train wheel) | Balance Rim Stop-Seconds | Magnetic Hold (electronic calibres) |
|---|---|---|---|
| Positioning resolution | 1 tooth (~6° at fourth wheel, 1 s) | Continuous, sub-second | Continuous, limited by IC clock |
| Amplitude impact during running | Zero if lever fully retracts | 5-15° loss if return spring weak | Zero — non-contact |
| Mechanism complexity | Moderate — yoke + lever + spring | Low — single sprung blade | High — drive electronics + coil |
| Typical service interval | 5-7 years (lever tip wear) | 5-7 years (spring fatigue) | Battery-life limited, 2-5 years |
| Cost to manufacture | Moderate — precision blade tip | Low — single stamping | High — IC and coil assembly |
| Best application fit | Workshop hacking, rate testing | Wristwatch seconds-sync | Quartz, tuning-fork, smart calibres |
| Failure mode | Lever tip wear, yoke pivot slop | Spring fatigue, rim scoring | IC failure, coil break |
Frequently Asked Questions About Geared Watch Stop
The return spring is too weak or the lever pivot is gummed with old oil, so the blade sits in light contact with the balance rim or fourth-wheel teeth even when the crown is pushed fully home. You can confirm this with a timing machine — measure amplitude with the movement assembled, then carefully lift the stop lever clear with a brass rod and re-measure. A 20°+ jump means the stop lever is dragging.
Fix by replacing the return spring or cleaning and re-oiling the lever pivot with a 9010-grade oil. Do not bend the existing spring to add tension — fatigued spring steel will snap within a few months.
For a 6497 specifically, balance-rim stop is the cleaner choice. The 6497's fourth wheel sits awkwardly under the train bridge with little real estate for a lever, while the balance is exposed on top of the movement. Most aftermarket hacking conversions for this calibre — including the well-known ones used in mil-spec homages — add a sprung wire that bears on the balance rim when the stem is pulled.
The trade-off is that a balance-rim stop has zero positioning resolution gear-wise — the wheel stops wherever it stops, with the seconds hand wherever it lands. For sub-second synchronisation you need a fourth-wheel engagement, which means a more involved bridge modification.
The stop lever is releasing the wheel before fully clearing the tooth. As the blade exits, it flicks the tooth flank and gives the wheel a tiny push — that push is exactly one tooth pitch, which on a 60-tooth fourth wheel is the 6° you're seeing.
The root cause is almost always lever geometry: the tip is too thick, or it's entering the gap at a slight angle because the yoke pivot has lateral play. Inspect the lever tip under 20× — a polished diagonal facet on one face confirms it. The fix is to dress the tip flat or fit a replacement, then check yoke pivot clearance against the spec, typically below 0.03 mm.
For a fully wound mainspring, yes — eventually. Residual mainspring torque keeps loading the contact between lever blade and tooth flank for as long as the watch is stopped. Over hours that's harmless. Over weeks of storage in the hacked position, the load can cold-flow the steel slightly at the contact point, leaving a witness mark on the tooth and a corresponding flat on the blade.
Best practice for long-term storage: let the mainspring run down before pulling the crown to the stop position, or simply leave the crown at position 1 and let the watch stop naturally when the spring runs out.
Almost always one of two causes. First, the train bridge was reseated with a slightly different end-shake on the fourth wheel — even 0.05 mm of extra vertical play shifts the tooth plane out of the lever's sweep arc, so the blade lands above or below the gap. Check end-shake with a feeler-style depth gauge against the original spec.
Second possibility is a yoke that was reinstalled rotated by one detent position, which puts the lever's rest position too far from the wheel for full travel. Pull the dial side and verify yoke geometry matches the technical sheet for the calibre.
For a 60-tooth fourth wheel engagement, the mechanism resolution is 1 second, but the operator-induced variation is usually larger. Reaction time on crown-release adds 100-300 ms of jitter, and the actual moment the lever clears the tooth lags the crown push by another 50-100 ms depending on yoke geometry.
Realistically expect ±0.5 second repeatability with a trained operator, ±1.5 seconds with an untrained one. To do better than that you need a balance-rim stop — which has finer resolution but no detent feel — or you need to abandon mechanical hacking entirely and use an electronic gating signal, as some observatory test fixtures do.
References & Further Reading
- Wikipedia contributors. Hacking (watchmaking). Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
