A worm-gear jumping motion with tumbler is an intermittent drive where a continuously rotating worm engages a partial-thread or single-lobe count wheel only briefly each revolution, then a spring-loaded tumbler arm snaps the wheel forward by one fixed step. Typical step rates run 20-200 advances per minute with positional repeatability inside ±0.5°. The mechanism converts smooth rotary input into discrete counts without clutches or solenoids, and you'll find it inside textile yardage meters, vintage taximeters, and Veeder-Root style mechanical totalisers.
Worm-gear Jumping Motion with Tumbler Interactive Calculator
Vary the cycle time, wheel tooth count, worm thread arc, and tumbler snap angle to see the intermittent index timing and jump geometry.
Equation Used
This calculator treats the mechanism as one indexed count per worm revolution. The wheel step is 360/N, the spring tumbler completes the final snap angle, and the worm supplies the remaining angular travel during the engaged thread arc.
- One count-wheel advance occurs per worm revolution.
- The worm drives part of the step, then the tumbler snaps the remaining angle to the stop.
- Thread arc is the engaged portion of one worm revolution.
- Snap angle is limited to the computed step angle.
The Worm-gear Jumping Motion with Tumbler in Action
The trick here is that the worm doesn't drive the count wheel continuously. Most of the worm's circumference is plain — turned smooth — with only one short helical thread segment cut into it. As the worm rotates, that segment engages a single tooth on the count wheel, pulls it part way forward, then disengages. At the moment of disengagement the spring-loaded tumbler arm — riding against a profiled cam face on the wheel — snaps over centre and finishes the indexing motion. This is the 'jumping' part. The wheel doesn't crawl to its rest position; it jumps the last few degrees under spring force and lands hard against a stop.
Why design it this way? Two reasons. First, the worm thread can be cut to engage briefly even when the input shaft is running at constant speed, so you don't need a clutch or a Geneva-style locking arc to get clean intermittent advance. Second, the tumbler hides the mesh tolerances. If the worm thread and wheel tooth had to mesh all the way to the rest position, any backlash would show up as count-wheel jitter and the digit reading would float. Letting the spring-loaded tumbler take over for the last 5-15° of travel means the wheel always lands on the same hard stop regardless of worm wear.
Get the tolerances wrong and the failure modes are predictable. If the worm thread is too long axially, the wheel over-rotates and the tumbler can't catch the next cam lobe — you skip counts. If the tumbler spring is too weak, the wheel stalls between worm disengagement and tumbler snap-over, and you get half-counts that read on one digit cycle but not the next. If the cam face on the wheel is worn flat, the tumbler loses its over-centre action entirely and the mechanism falls back to whatever residual motion the worm imparted — usually about half a step short.
Key Components
- Single-segment worm: A worm shaft with most of the circumference turned smooth and only one helical thread segment, typically subtending 30-60° of shaft rotation. The thread engages the count wheel tooth for a fraction of each input revolution. Thread profile is usually a standard 14.5° or 20° pressure angle ground to ±0.05 mm pitch tolerance.
- Count wheel with cam profile: The driven wheel carries 10, 12, or sometimes 60 teeth around its rim and a matching cam profile on its side face. Each tooth meshes briefly with the worm; each cam lobe gives the tumbler its over-centre kick. Tooth-to-tooth indexing accuracy of ±0.3° is normal on a well-built unit.
- Spring-loaded tumbler arm: A pivoted arm pressed against the wheel's cam face by a torsion or compression spring providing 0.5-2.0 N at the contact point. The arm snaps over each cam lobe to complete the index step and hold the wheel firmly against its rest position between counts.
- Hard stop or detent: A fixed pin or shoulder the count wheel lands against at the end of each jump. This is what kills the inherent backlash — the tumbler spring force loads the wheel against this stop, so the readout digit position is always referenced to the stop, not to worm-gear backlash.
- Input shaft and bearing pair: Carries the worm at constant input speed, typically 20-300 RPM in counter applications. Bearing radial play above 0.05 mm lets the worm walk axially during engagement and causes inconsistent step length — a common cause of 'soft' counts in worn units.
Who Uses the Worm-gear Jumping Motion with Tumbler
The worm-gear jumping motion with tumbler shows up wherever a continuously running shaft has to produce one clean count per revolution without electronics. It's a 19th and early-20th century solution that survives because it's cheap, silent, and self-locking. You'll still find it in refurbished metering equipment, heritage textile machinery, and demonstration models in engineering schools.
- Textile manufacturing: Yardage meters on warp-beaming frames at heritage mills like the American Textile History Museum's working Crompton & Knowles loom, where a worm-and-tumbler totaliser logs woven yardage off the take-up roll.
- Transportation metering: Mechanical taximeters of the Argo and Halda eras, where a flexible drive shaft from the gearbox turns a worm that advances the fare digit wheel by one count every set distance.
- Industrial flow metering: Veeder-Root style mechanical totalisers paired with positive-displacement liquid flow meters in legacy fuel-dispenser pump heads, registering each shaft revolution as one volumetric count.
- Agricultural machinery: Acreage counters on vintage John Deere and Massey-Harris seed drills, where a ground wheel turns the worm and the tumbler indexes a totaliser logging seeded acres.
- Process equipment: Stroke counters on reciprocating dosing pumps such as ProMinent Vario plunger pumps retrofitted with mechanical totalisers, where each pump cycle drives the worm one revolution.
- Educational demonstration: Working sectional models of intermittent mechanisms at engineering departments like the Cornell Reuleaux Collection, where the worm-and-tumbler unit illustrates slip-and-snap kinematics for undergraduate kinematics classes.
The Formula Behind the Worm-gear Jumping Motion with Tumbler
The number you most often need is the count rate — how many discrete advances per minute you get from a given worm input speed and how that scales with count-wheel tooth count. At the low end of the typical range (around 20 RPM input) the mechanism feels deliberate and you can watch each jump individually. At the nominal mid-range (60-100 RPM) the tumbler snaps cleanly and the digit changes look crisp. Push past the high end (250-300 RPM) and the tumbler can't reset between engagements — the spring rate becomes the limiting factor, not the gear math.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Crate | Count advance rate at the wheel | counts/min | counts/min |
| Nworm | Worm input shaft speed | RPM | RPM |
| Sworm | Number of engagement segments cut into the worm per revolution (usually 1) | segments/rev | segments/rev |
| Zwheel | Engagement-to-step ratio (set to 1 for one count per worm rev, higher if the wheel needs multiple worm revs per count) | dimensionless | dimensionless |
Worked Example: Worm-gear Jumping Motion with Tumbler in a vintage taximeter restoration shop
A vintage taximeter restoration shop in Berlin is rebuilding a 1962 Halda Taxameter for a private collector. The drive cable runs from the gearbox tail to a worm shaft inside the meter head, and the worm advances a 10-tooth fare count wheel through a tumbler-armed jumping motion. The shop needs to know the count rate at three operating speeds — 30 RPM (slow city crawl), 90 RPM (nominal cruise) and 240 RPM (autobahn return run) — to confirm the fare-digit cadence matches the original Halda factory test sheet.
Given
- Sworm = 1 segment/rev
- Zwheel = 1 dimensionless
- Nworm,low = 30 RPM
- Nworm,nom = 90 RPM
- Nworm,high = 240 RPM
Solution
Step 1 — at nominal 90 RPM, plug the values straight in:
That's 1.5 counts per second — fast enough that the digit wheel looks like it's incrementing continuously to a passenger glancing at the meter, but slow enough that the tumbler has roughly 670 ms to snap over and reset between jumps. Comfortably inside the spring's reset window.
Step 2 — at the low end of the operating range, 30 RPM:
One count every 2 seconds. You can clearly see the individual jumps and hear each tumbler snap. This is what the taxi feels like creeping through Friedrichstraße traffic — the digit advances are deliberate and visible.
Step 3 — at the high end, 240 RPM:
4 counts per second, with only 250 ms between snaps. On the original Halda spec sheet this is right at the published ceiling. In practice, if the tumbler torsion spring has aged below its original 1.2 N tip force, the wheel won't fully seat against the hard stop before the next worm engagement, and you start dropping or smearing counts. Above ~280 RPM no original-spec spring will keep up.
Result
Nominal count rate is 90 counts/min, matching the Halda factory test cadence within the published ±2% tolerance. The 30/90/240 RPM range tells you the sweet spot sits between roughly 60 and 150 RPM — below 60 you can hear individual snaps which the original engineers considered acceptable but not refined, and above 150 you start eating into the tumbler's reset margin. If your rebuilt unit measures 75 counts/min at a verified 90 RPM input, the most likely culprits are: (1) a worm thread engagement segment that has been re-cut shorter than the 45° factory spec, shortening each pull-in arc, (2) a count-wheel tooth chipped at the engagement face, causing the worm to slip past on every third or fourth rev, or (3) drive-cable lost motion above 3° at the meter input, which subtracts directly from worm rotation per gearbox revolution.
Worm-gear Jumping Motion with Tumbler vs Alternatives
The worm-and-tumbler isn't the only way to get one count per input revolution. Geneva drives and ratchet-and-pawl mechanisms cover the same intent. Each has its own engineering envelope — here's how they compare on the dimensions a designer actually cares about.
| Property | Worm-gear jumping with tumbler | Geneva drive | Ratchet and pawl |
|---|---|---|---|
| Practical max count rate | ~280 counts/min | ~600 counts/min | ~400 counts/min |
| Indexing accuracy | ±0.3° at the wheel | ±0.05° (locked by Geneva geometry) | ±0.5° (pawl seating dependent) |
| Backlash sensitivity | Low — tumbler hard-stops the wheel | Very low — geometric lock | Medium — pawl tip wear opens up over time |
| Build cost (small batch) | Low — one segmented worm + spring | High — precision Geneva slot grinding | Low — stamped pawl and ratchet |
| Service life before refurb | ~10 million counts before tumbler spring fatigue | ~50 million indexes before slot wear | ~5 million counts before pawl tip rounding |
| Audible noise | Soft snap each count | Quiet — pure rolling | Sharp click each count |
| Best application fit | Continuous-shaft mechanical counters | High-precision indexing tables | Manually pumped or low-cycle counters |
Frequently Asked Questions About Worm-gear Jumping Motion with Tumbler
That's tumbler over-travel. When the input cable transmits a sudden torque spike — common with kinked or freshly relubed flexible drive cables that release stiction abruptly — the worm rotates further during the brief engagement window than the design assumed. The wheel gets pushed past one cam lobe and the tumbler snaps over the next one too.
Fix the cable behaviour first (replace if the inner core feels notchy when you rotate it slowly), and verify the tumbler spring preload. A spring that has been over-stretched during reassembly will let the arm bounce off the cam face instead of damping the motion at the rest position.
Multi-start worms give you 2 or 3 counts per input revolution but each engagement window shrinks proportionally. With a 2-start worm the tumbler has half the reset time, so you have to either stiffen the tumbler spring (typically 60-80% increase in preload) or accept a lower max input RPM. The math doesn't lie — total tumbler events per second is what governs the spring's duty.
Rule of thumb: if you need under 200 counts/min, stay single-start. If you need 200-500 counts/min and your input shaft can't spin faster, go 2-start and upgrade the spring. Above 500 counts/min the worm-and-tumbler is the wrong mechanism — switch to a Geneva drive.
Check the input coupling for elastic wind-up before you blame the mechanism. On flexible-cable inputs (Halda taximeters, ground-wheel acreage counters) the cable twists under load and recovers between engagements, so the worm sees less rotation than the input shaft over a measurement window. An 8% deficit is consistent with about 30° of cable wind-up per revolution at typical input torques.
Diagnostic check: paint a witness mark on both ends of the input cable and run the unit at the test speed. If the marks lag each other by more than 5° at steady state, the cable is your loss path, not the gear.
Chatter at seating means the spring force is high relative to the damping, so the arm bounces off the hard stop. Three common causes: (1) someone replaced the original tumbler spring with a stiffer aftermarket part, (2) the rest-position pad has hardened with age and lost its damping property → original pads were usually a fibre or leather composite, and (3) the cam profile has been polished too aggressively during a rebuild, removing the deceleration ramp that should slow the arm before contact.
If the original cam had a 5-10° deceleration radius and yours is now flat, the arm hits the stop at full velocity. That's a re-machining job, not a spring swap.
Not without redesigning the worm thread and the tumbler cam. The single-segment worm thread has a defined engagement direction — the helix angle pulls the wheel one way, and reversing the worm just slips the thread off the tooth without imparting motion. Likewise, the cam profile under the tumbler is asymmetric: the lead-in ramp on one side, the over-centre snap on the other.
For bidirectional counters you either use two separate worm-and-tumbler stages with selectable engagement (mechanically complex), or switch to a ratchet-and-pawl with a reversible pawl, or bite the bullet and use an electronic counter. Most heritage bidirectional counters chose the third option once it became affordable.
Counter-intuitive but real. Higher preload means higher snap-over force, which means higher impact velocity at the hard stop. Above a certain preload the wheel rebounds off the stop by 1-3° before settling, and the readout digit floats between two positions until you tap the case.
The factory spring is usually tuned to land the wheel at roughly 70% of the maximum preload that would cause rebound — that gives you reset margin without bounce. If you've increased preload to fix a sluggish snap, check whether you've actually introduced rebound by watching the wheel under stroboscopic light at operating speed.
References & Further Reading
- Wikipedia contributors. Intermittent mechanism. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
