An electric escapement is a mechanism that releases one tooth of an escape wheel per electrical pulse, converting discrete solenoid or coil firings into precise incremental rotation. Matthias Hipp built the most famous early version in 1842 with his toggle-driven master clock, which drove dozens of slave dials from a single pendulum. The pulse energises a coil, the coil retracts a pallet or pawl, the wheel advances exactly one tooth, and the pallet drops back to lock. That single-tooth release is what gives ticket machines, slave clocks, and step indexers their repeatable motion without a continuous drive.
Electric Escapement Interactive Calculator
Vary escape-wheel tooth count, pulse count, and pulse rate to see the one-tooth-per-pulse angular indexing motion.
Equation Used
The escapement advances one wheel tooth for each electrical pulse. With N teeth, one tooth pitch is 360/N degrees, so p pulses rotate the wheel by p x 360/N degrees. A steady pulse rate f gives wheel speed 60f/N rpm.
- Each electrical pulse releases exactly one tooth.
- No missed steps or double-feeds occur.
- Pulse rate is steady over the calculation interval.
How the Electric Escapement Actually Works
The core trick of any electric escapement is that it converts an electrical event into a mechanical event of fixed angular size. You send a pulse — typically 12 V or 24 V at 50 to 200 ms — into a solenoid coil. The coil pulls an armature, the armature is mechanically linked to a pallet or pawl that normally blocks the escape wheel, and when it lifts, the wheel rotates by exactly one tooth pitch under the load of a spring, weight, or stepper torque. Drop the pallet back, the next tooth catches, and you are locked again until the next pulse. One pulse, one tooth. That is the entire promise.
Get the geometry wrong and the promise breaks fast. The pallet lift must exceed tooth height by at least 0.2 mm for reliable release — too shallow and the wheel skips intermittently when bearing drag varies. The pallet must drop back BEFORE the next tooth has rotated past the catch face, so the coil release time has to be shorter than the wheel's tooth-to-tooth travel time. On a Hipp toggle master clock running 1 pulse per second, that constraint is trivial. On a high-speed ticket dispenser firing at 20 Hz, sloppy coil decay can let the wheel free-wheel two teeth in a single pulse — the classic double-feed failure that ruins a printed ticket.
The other failure mode you see in the field is impulse-energy mismatch. If the coil's pull force is just barely enough to lift the pallet against its return spring, dust, dried grease, or a weakened return spring will silently drop you below the release threshold. Symptoms are missed counts on a step indexer or a slave clock that loses one minute every few hours. Always size the solenoid for at least 1.5× the static lift force, and pick a return spring stiff enough to reseat the pallet in under half the pulse period.
Key Components
- Escape Wheel: Toothed wheel that advances one tooth per electrical pulse. Tooth count typically ranges from 15 to 60 — a 30-tooth wheel gives 12° per pulse. Tooth profile is usually a club or ratchet form with a 5° to 15° impulse face so the pallet receives a clean push back as it drops, helping the armature reset.
- Solenoid Coil: Electromagnet that pulls the armature when energised. Common spec is 12-24 V DC, 0.5-3 W continuous rating, with pulse currents of 100-500 mA. Coil resistance and inductance set the rise time — a 30 mH coil at 24 V hits useful pull force in about 5 ms, which is the practical lower bound for pulse rate.
- Pallet or Pawl: Hardened lever that blocks the escape wheel between pulses. Tip hardness should be 58-62 HRC to resist tooth wear. The geometric constraint is unforgiving — pallet lift must clear tooth height by 0.2 mm minimum, and the locking face must sit at 1° to 3° draw angle so wheel torque pulls the pallet INTO engagement rather than out of it.
- Return Spring: Resets the pallet after the coil de-energises. Spring force has to overcome armature inertia and any residual magnetism in the core within roughly half the pulse-to-pulse interval. Too weak and you get drop-out misses; too strong and the coil cannot lift it reliably at end-of-life voltage.
- Drive Source: Provides the torque that rotates the wheel during the unlocked moment. Could be a falling weight (Hipp clocks, ~0.05 N·m typical), a mainspring, a constant-torque motor, or in modern systems a stepper held in detent. Drive torque must be high enough to overcome bearing drag plus the impulse-face friction in well under the coil pulse duration.
- Contact or Trigger Circuit: Generates the pulse. In an old Hipp master clock this was a mercury-tilt contact closed once per pendulum swing; in modern equipment it is a microcontroller output through a MOSFET with a flyback diode across the coil. The flyback diode is mandatory — without it the coil's collapsing field will destroy the driver transistor within hours.
Real-World Applications of the Electric Escapement
Electric escapements show up wherever you need to convert counted electrical pulses into counted mechanical steps without a feedback loop. They are simple, cheap, and self-locking between pulses, which is why they survived in industrial equipment long after stepper motors became common. You will find them in timekeeping, ticketing, vending, and any low-cost indexer where a stepper is overkill but a continuous motor cannot give you repeatable position.
- Horology: Hipp toggle master clocks at the Royal Observatory in Neuchâtel, Switzerland, drove slave dials across the city using a single solenoid-pulsed escapement firing once per second.
- Public Transport Ticketing: Almex and Setright bus ticket machines used a sprung escapement released by a hand-crank-triggered electric contact to advance the paper roll exactly one ticket length.
- Industrial Counting: Veeder-Root mechanical counters with electric reset use a solenoid-driven escapement to advance the units wheel one digit per input pulse.
- Vending and Coin Dispensing: National Rejectors coin hoppers used an electric escapement to release one coin per pulse from a stacked column, eliminating the need for an encoder.
- Telecommunications Switching: Strowger step-by-step telephone exchanges used solenoid-driven escapement-style ratchets to advance selector wipers one contact per dialled pulse, the basis of automatic telephony from the 1890s through the 1980s.
- Scientific Instruments: Chart recorders such as the Esterline-Angus AW used escapement-driven paper advance for low-power, battery-operated field data logging.
The Formula Behind the Electric Escapement
The useful design equation links pulse rate, escape wheel tooth count, and output shaft speed. What you really want to know is: for a given pulse frequency, how fast does the output shaft turn, and where does the mechanism start to fall apart? At the low end of practical pulse rates (1 Hz, like a seconds clock) the escapement is loafing — bearing drag and dust are your only enemies. At the high end (above about 20 Hz for typical small-coil escapements) coil thermal load and pallet-bounce time start eating into reliability. The sweet spot for most industrial work is 2 to 10 Hz.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| ωout | Output shaft angular speed | °/s | °/s |
| Nteeth | Number of teeth on the escape wheel | teeth | teeth |
| fpulse | Electrical pulse frequency driving the coil | Hz | pulses/s |
| θstep | Angle advanced per pulse (= 360° / Nteeth) | ° | ° |
Worked Example: Electric Escapement in a museum ticket dispenser
Restoring a 1965 Almex Model E bus ticket dispenser for a tram museum exhibit. The escape wheel has 24 teeth, the coil runs from a 24 V DC supply with 80 ms pulses triggered by the operator's hand lever, and you need to confirm the paper feed advances one ticket (15° of escape wheel rotation per ticket) per lever pull, with the system robust enough to handle a determined visitor cranking 4 tickets per second.
Given
- Nteeth = 24 teeth
- θstep required = 15 ° per ticket
- Vcoil = 24 V DC
- Pulse width = 80 ms
- fpulse,max = 4 Hz
Solution
Step 1 — confirm the geometry. With 24 teeth on the escape wheel, the angle per pulse is:
That matches the required 15° per ticket exactly. One pulse, one ticket. No gearing needed between the escape wheel and the paper-feed roller.
Step 2 — compute output speed at the nominal duty rate. A typical museum visitor pulls the lever at about 1 Hz:
At 15°/s the paper crawls forward at a clean, deliberate pace — the visitor can actually see one ticket emerge per pull, which is exactly the tactile feedback the original Almex design intended.
Step 3 — check the low-end and high-end of the operating range. At 0.2 Hz (a hesitant tourist taking five seconds between pulls):
That is fine — the escapement does not care how slow you go, only that each pulse is fully formed. At the 4 Hz upper limit:
Sixty degrees per second sounds modest, but the practical problem is the 80 ms pulse width itself. At 4 Hz the pulse-to-pulse interval is 250 ms, so the coil is energised for 80 ms and the pallet has only 170 ms to drop, reseat, and lock the next tooth. With a soft return spring or any residual core magnetism, you will see the wheel free-wheel and dispense two tickets on a single pull. The Almex design solves this with a stiff phosphor-bronze return spring sized for a 40 ms reset.
Result
At nominal 1 Hz operator cadence the paper feed advances 15°/s, exactly one ticket per lever pull as designed. Across the operating range, behaviour scales linearly from 3°/s at a slow 0.2 Hz pull rate up to 60°/s at the 4 Hz mechanical limit — but the sweet spot is below about 2.5 Hz, above which pulse-decay time starts eating reset margin. If you measure double-tickets being dispensed on rapid pulls, the three failure modes to check in order are: (1) missing or undersized flyback diode across the coil, letting the field collapse slowly and holding the pallet up too long; (2) a return spring weakened by 60 years of set, no longer reseating in under 100 ms; (3) hardened grease in the pallet pivot dragging the reset stroke past the next-tooth catch window. Clean the pivot, fit a 1N4007 across the coil, and replace the spring before you blame the geometry.
When to Use a Electric Escapement and When Not To
Electric escapements compete with stepper motors and ratchet-and-pawl mechanisms whenever you need pulse-counted incremental motion. The choice usually comes down to cost, pulse rate, and whether you need closed-loop position feedback. Here is how they stack up on the dimensions that matter for a real build.
| Property | Electric Escapement | Stepper Motor | Solenoid Ratchet & Pawl |
|---|---|---|---|
| Max practical pulse rate | 1-20 Hz | 200-2000 full steps/s | 1-10 Hz |
| Position accuracy per step | ±0.5° (set by tooth pitch) | ±5% of step angle (typ ±0.09°) | ±1° (depends on pawl seating) |
| Cost per unit (small batch) | $8-25 | $15-60 plus driver | $5-15 |
| Holding torque between pulses | Mechanical lock, infinite holding | Detent torque only unless powered | Mechanical lock via pawl |
| Power consumption when idle | 0 W (purely mechanical lock) | 1-10 W if energised for hold | 0 W |
| Lifespan (cycles to wear-out) | 10⁶-10⁷ cycles | 10⁸+ cycles | 10⁵-10⁶ cycles |
| Best application fit | Counters, ticket dispensers, slave clocks | CNC, robotics, precision indexing | Cheap one-shot advance |
Frequently Asked Questions About Electric Escapement
You are almost certainly running into pulse-overlap. At higher pulse rates the coil's inductance prevents the magnetic field from collapsing fast enough between pulses, so the pallet never fully reseats before the next pulse arrives. The wheel sees a partially-lifted pallet and slips past it.
Quick check: scope the voltage across the coil and look at the decay tail. If the field is still above 30% of peak when the next pulse fires, you need a faster decay. Add a Zener diode (typically 36-50 V) in series with your flyback diode — this clamps the kickback at a higher voltage and forces the current to decay roughly 5-10× faster than a plain diode.
Neither will give you 1° directly — 60 teeth gives 6° per pulse and 30 gives 12°. To hit 1° you need to gear down 6:1 from a 60-tooth wheel, or rethink the design. For sub-degree resolution a stepper motor is almost always the right answer; escapements top out around 60 teeth before tooth strength becomes marginal at typical small-mechanism scales (sub-30 mm escape wheel diameter).
If you must use an escapement, use 60 teeth, a 6:1 reduction gear, and accept that backlash in the gear train will eat about half your theoretical resolution.
Run a slow-motion test. Drop the pulse rate to 0.5 Hz and energise manually with a pushbutton. If every press still produces exactly one tooth advance, the mechanics are fine and your problem is electrical — likely either insufficient pulse width, sagging supply voltage under load, or a marginal flyback path letting the coil hold the pallet too long.
If even slow manual pulses sometimes miss, the issue is mechanical: pallet lift below the 0.2 mm clearance threshold, return spring fatigue, or contamination in the pivot. Pull the pallet, inspect the tip for rounding (worn pallets lose draw angle and slip under wheel torque), and check the lift geometry with feeler gauges.
Three culprits, in order of likelihood. First, mounting orientation — if you bench-tested horizontal and the final mount is vertical, gravity now adds or subtracts from the pallet return-spring force. The same spring that reseats the pallet in 40 ms horizontal might take 80 ms fighting gravity, pushing you over the pulse-overlap cliff.
Second, supply impedance. Bench supplies have stiff regulation; an in-product 24 V rail shared with other loads can sag 2-4 V during the coil pull, dropping coil force below the lift threshold. Third, magnetic coupling — mounting the coil close to a steel chassis changes its effective inductance and pull force by 10-20%. Always characterise the coil in the actual mounting geometry, not free-air.
For a working museum piece, almost never. The Hipp toggle has a specific click-and-pause cadence that is part of the artefact's character, and a stepper running open-loop in micro-step mode will never replicate that mechanical signature. Restore the original — the failure modes are well-understood (worn pallet tips, oxidised mercury contacts, weak pendulum impulse) and parts can be machined or sourced from horological suppliers.
If the clock is purely functional and authenticity does not matter, a stepper plus an RTC gives better long-term accuracy (a stepper-driven clock holds time to the crystal, typically ±20 ppm; a Hipp master clock drifts with temperature and pendulum amplitude, typically ±1-5 seconds per day).
The pulse must be long enough for two events: the coil current to rise to its pull-force threshold, AND the armature to physically move through the lift stroke. For a typical small escapement with a 30 mH coil and 2-3 mm armature travel, the rise time is around 5 ms and the mechanical travel takes another 10-20 ms. That sets a hard lower bound around 25 ms.
Practical advice: start at 50 ms and shorten while watching for missed counts. If you go below 30 ms and still need shorter pulses, you need a lower-inductance coil (fewer turns, higher current) — not a shorter pulse on the existing coil.
References & Further Reading
- Wikipedia contributors. Electromechanical clock. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
