A modified worm-and-pinion is a hybrid gear pair that replaces the conventional cylindrical worm with a profile-corrected variant — typically a globoid (hourglass) worm or a duplex worm with shifted lead angles — meshing with a throated wheel that wraps around it. The shape change spreads contact across more teeth at once, which solves the point-contact wear and backlash problem of a standard worm drive. You get reduction ratios from 20:1 up to 300:1 in a single stage with measurable backlash under 1 arc-minute, which is why you find them on the AD Series Cone Drive azimuth heads steering 12 m radio dishes.
Worm-and-pinion Modified Interactive Calculator
Vary input torque, speed, ratio, efficiency, and contact teeth to see torque multiplication, speed reduction, heat loss, and load sharing in a globoid worm mesh.
Equation Used
The calculator applies the worm reduction ratio and overall efficiency to estimate output torque, output speed, and lost power. The contact-teeth input represents the modified globoid/throated mesh spreading load across several teeth rather than a single point contact.
FIRGELLI Automations - Interactive Mechanism Calculators
- Overall efficiency represents mesh, bearing, and lubricant losses.
- Ratio is single-stage worm reduction from input worm to output wheel.
- Contact teeth share transmitted load equally across the modified mesh band.
- Back-driving and self-locking limits are not evaluated.
How the Worm-and-pinion (modified) Actually Works
A standard worm-and-pinion (which engineers more accurately call a worm and worm wheel) drives torque from a screw-shaped worm into a helical gear at 90°. The contact patch is small — basically a point that sweeps along the wheel tooth — and that's the source of every problem the modified versions exist to solve. Localised contact stress means high Hertzian pressure, sliding friction, heat, and backlash that grows fast as the wheel face wears. A modified worm-and-pinion attacks this by changing either the worm profile, the wheel profile, or both, so the contact patch becomes a line or a curved band sharing load across 3 to 5 teeth simultaneously.
The two common modifications you'll see in industry are the globoid worm (also called an hourglass worm or throated worm) and the duplex worm. A globoid worm is concave along its length, so it physically wraps around the wheel — the throated worm wheel and throated worm together form a double-enveloping pair. Cone Drive in Traverse City has built their business on this geometry since 1925. A duplex worm uses two different lead angles on opposite flanks of the same thread, so axial shift of the worm shaft tightens or loosens mesh on demand. That's how you get an anti-backlash worm drive without splitting the wheel.
If the centre distance is off by more than 0.05 mm on a globoid pair, the contact pattern shifts off the throat and you lose half your load-sharing teeth — efficiency drops from 90% to roughly 70% and the wheel face starts pitting within a few hundred hours. If the duplex worm shift is over-tightened past about 30 arc-seconds of preload, you'll cook the bronze wheel from sliding heat at any duty cycle above 40%. The most common failure modes are wheel-tooth scuffing from lubricant breakdown, throat wear from misalignment, and worm thread pitting if the steel worm runs harder than 58 HRC against an undersized bronze rim.
Key Components
- Globoid (Hourglass) Worm: A steel worm with a concave longitudinal profile that wraps around the wheel's outside diameter. Typical case-hardened 8620 steel ground to 58-62 HRC, with thread profile tolerance held to ±0.005 mm on lead. The concavity engages 3 to 5 wheel teeth simultaneously instead of the single point a cylindrical worm produces.
- Throated Worm Wheel: Bronze (usually SAE 65 or aluminium-bronze) gear with a concave face that mirrors the worm's curvature. The throat radius matches the worm's pitch radius within 0.025 mm. This is the wear part — bronze sacrifices itself to protect the steel worm, and you'll typically see 8,000 to 20,000 hours service life depending on duty cycle.
- Duplex Worm Variant: An alternative modification where one cylindrical worm carries two different lead angles — say 4° on the leading flank and 4°15' on the trailing flank. Axial position of the worm shaft becomes the backlash adjustment. Holding shift tolerance to ±0.01 mm gets you backlash under 30 arc-seconds without splitting the wheel.
- Centre-Distance Adjuster: Eccentric bushings or shimmed bearing carriers that let you set centre distance to within 0.05 mm at assembly. On Cone Drive's MHU-series gearboxes the adjuster is a graduated eccentric — one click equals 0.012 mm of centre-distance change.
- Thrust Bearing Pair: Angular-contact bearings on the worm shaft that take the axial reaction load — which on a high-ratio worm drive can exceed input torque by 10× to 50×. Undersizing here is the silent killer of modified worm drives. Standard practice uses paired 7000-series ABEC-5 bearings preloaded to 200-400 N.
Real-World Applications of the Worm-and-pinion (modified)
Modified worm-and-pinion drives show up wherever you need huge reduction ratios in one stage, self-locking behaviour for safety, and tighter backlash than a vanilla worm pair can deliver. The trade-off is always efficiency — you'll typically see 70% to 92% efficiency depending on lead angle and lubrication — but in applications where the load doesn't need to back-drive, that efficiency penalty is the feature, not the bug.
- Radio Astronomy: Cone Drive AD Series azimuth and elevation heads on 12 m to 25 m radio telescope dishes, including units serving the Very Long Baseline Array stations.
- Solar Tracking: Kinematics Manufacturing SE-series slewing drives on utility-scale solar tracker installations — a single SE21 unit positions a 40 kW tracker row to ±0.1°.
- Marine: Steering gear quadrant drives on tugboats and small ferries, where self-locking behaviour holds the rudder against wave loads with the hydraulic ram unloaded.
- Industrial Valve Actuation: Rotork IW-range manual override gearboxes on pipeline ball valves up to DN800, providing 270:1 reduction so a single operator can close a valve seeing 200 bar of line pressure.
- Stage and Theatre: ETC and JR Clancy automated rigging winches in performing-arts venues, where the self-locking globoid drive holds scenery loads if power is lost mid-cue.
- Machine Tool Indexing: Nikken CNC rotary tables — the 5AX-201 series uses a duplex worm modification to hold ±15 arc-seconds positioning accuracy across the full 360°.
The Formula Behind the Worm-and-pinion (modified)
The single number that drives most modified worm-and-pinion design decisions is the gear ratio, which sets your output torque and output speed simultaneously. For a globoid pair, the effective ratio also depends on how many teeth are in mesh under load — at the low end of typical operating range (around 20:1) you're effectively running a high-lead worm with 4° to 6° lead angle, getting 88% to 92% efficiency but only marginal self-locking behaviour. At the high end (200:1 to 300:1) the lead angle drops below 2° and the drive becomes firmly self-locking, but efficiency falls to 50%-70% and heat dissipation becomes the limiting factor. The sweet spot for most industrial applications sits at 40:1 to 80:1 where you keep efficiency above 80% while staying self-locking under static load.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| i | Gear ratio (input revs per output rev) | dimensionless | dimensionless |
| Nwheel | Number of teeth on the worm wheel | teeth | teeth |
| Zworm | Number of starts (threads) on the worm | starts | starts |
| Tout | Output torque at the wheel shaft | N·m | lb-ft |
| Tin | Input torque at the worm shaft | N·m | lb-ft |
| η | Mesh efficiency (0.50 to 0.92 typical) | dimensionless | dimensionless |
Worked Example: Worm-and-pinion (modified) in a Surface-Grinder Cross-Feed Drive
Sizing the modified globoid worm-and-pinion driving the cross-feed handwheel-to-leadscrew reduction on an Okamoto ACC-1224EX surface grinder rebuild, where the operator's handwheel must produce 0.005 mm of table movement per click of the indexing detent and the leadscrew has a 4 mm pitch. The handwheel detent gives 100 clicks per revolution and the input torque available at the handwheel is 1.2 N·m. The modified globoid pair must deliver enough reduction so each click moves the leadscrew exactly 0.005 mm, while keeping output torque high enough to overcome 18 N·m of cross-feed friction at the table.
Given
- Pitchleadscrew = 4 mm/rev
- Δxper-click = 0.005 mm
- Clicksper-rev = 100 clicks
- Tin = 1.2 N·m
- Tfriction = 18 N·m
- ηnominal = 0.82 —
Solution
Step 1 — work out how many leadscrew revs we need per handwheel rev. Each click is 0.005 mm and there are 100 clicks per handwheel rev, so the handwheel produces 0.5 mm of table travel per rev. The leadscrew advances the table 4 mm per rev:
That's the nominal reduction the worm-and-pinion must deliver. Step 2 — pick the worm and wheel tooth counts. An 8:1 ratio is at the low end of the typical modified-worm operating range, so we'd use a 2-start worm with a 16-tooth throated wheel:
Step 3 — compute output torque at nominal 82% efficiency:
That's well below the 18 N·m we need to overcome cross-feed friction, so the 8:1 single-stage choice fails the torque check. The handwheel feels light but the table won't budge cleanly. Step 4 — re-size at the high end of the typical range. Push to a 1-start worm with a 40-tooth wheel for 40:1, accepting efficiency drop to roughly 65% because the lead angle falls below 2.5°:
Now we clear the friction load with margin, but each handwheel click only moves the table 0.001 mm — five times finer than the operator wants, so feeds become tediously slow. Step 5 — at the low-end of the practical compromise, a 16:1 ratio (1-start worm, 16-tooth wheel) at 78% efficiency:
Still 17% short of the friction torque. The clean answer is a 24:1 ratio (1-start worm, 24-tooth wheel) at roughly 72% efficiency, giving 20.7 N·m output and 0.00167 mm per click — tight enough for grinding work, with adequate torque headroom.
Result
The nominal sized solution is a 24:1 modified globoid worm-and-pinion delivering 20. 7 N·m output torque with 0.00167 mm of table travel per handwheel click. In practice that feels like firm but smooth handwheel resistance under the operator's fingers — not stiff, not loose — and the table moves in clean, repeatable increments without stick-slip. Across the operating range, the 8:1 design produces light handwheel feel but stalls under load, the 24:1 sweet spot balances both, and the 40:1 design feels heavy and slow but holds position perfectly when locked. If your built unit measures less than 18 N·m output, the three most likely causes are: (1) thrust-bearing preload set above 600 N, which adds 3-5 N·m of parasitic drag at the worm; (2) wrong lubricant — running ISO VG 220 mineral oil instead of the specified ISO VG 460 polyglycol drops mesh efficiency by 8-10 percentage points on a low-lead-angle pair; or (3) bronze wheel hardness above 110 HB, which the standard SAE 65 bronze should not exceed and which signals the wrong alloy was supplied.
Worm-and-pinion (modified) vs Alternatives
Modified worm-and-pinion drives compete with planetary gearboxes, harmonic drives, and standard cylindrical worm pairs in the high-reduction single-stage market. The decision usually comes down to backlash budget, efficiency tolerance, and unit cost — these are the three axes where the technologies separate cleanly.
| Property | Modified Worm-and-Pinion | Standard Worm-and-Wheel | Harmonic Drive |
|---|---|---|---|
| Single-stage reduction ratio | 20:1 to 300:1 | 5:1 to 100:1 | 30:1 to 320:1 |
| Backlash (arc-min) | 0.5 to 2 | 5 to 30 | less than 0.5 |
| Efficiency at full load | 65% to 92% | 55% to 88% | 70% to 85% |
| Self-locking under static load | Yes (below 4° lead angle) | Yes (below 5° lead angle) | No |
| Typical service life (hours) | 8,000 to 20,000 | 5,000 to 15,000 | 10,000 to 35,000 |
| Relative cost (single unit) | 1.0× baseline | 0.4× baseline | 3× to 6× baseline |
| Shock load capacity | High — 3× nominal | Medium — 2× nominal | Low — 1.5× nominal |
| Best application fit | Valve actuators, slew drives, rigging | General industrial reduction | Robotics, precision indexing |
Frequently Asked Questions About Worm-and-pinion (modified)
That's the trade-off you signed up for when you chose the globoid geometry. The modified worm engages 3 to 5 teeth simultaneously instead of one, which spreads load but also multiplies the sliding contact area producing friction heat. At identical input power you'll see 15-25% more case temperature on the globoid pair.
If your case is running above 80°C steady-state, check that the lubricant is rated for the higher contact temperature — ISO VG 460 polyglycol is the standard for globoid drives, and substituting a mineral gear oil will drop your heat-removal capacity by roughly 30%. Add forced-air cooling on the case before you blame the gearset.
Pick based on whether you need adjustability or absolute load capacity. A duplex worm gives you axial-shift backlash adjustment that you can re-tune in the field as the wheel wears — Nikken and Tsudakoma rotary tables use this so an end-user can restore positioning accuracy after 10,000+ hours. The trade-off is the duplex worm is still a cylindrical worm, so you keep the line-contact patch, not the wider envelope contact a globoid offers.
A globoid pair gives you 2-3× the load capacity at the same envelope size and tighter as-built backlash, but adjustment requires re-shimming centre distance and you can't field-tune it once installed. Rule of thumb: indexers and machine-tool tables go duplex, slew drives and valve actuators go globoid.
Almost always one of two things: thrust-bearing end-float, or worm shaft deflection under reverse load. The factory measures backlash with the worm shaft fully constrained axially. If your installation has 0.05 mm of axial slop in the thrust stack, that translates directly through the lead angle into 3-5 arc-minutes of apparent backlash at the wheel.
Check the worm-shaft thrust preload first — you want zero axial play and around 200-400 N of preload on the angular-contact pair. If the bearings are fine, dial-indicate the worm shaft for radial deflection while applying reverse torque at the wheel. More than 0.02 mm of deflection means the worm shaft is undersized or the bearing span is too long.
Not reliably. Self-locking on a worm drive depends on lead angle staying below the friction angle of the mesh — typically around 4° for steel-on-bronze with good lubrication. If you try to back-drive a 2° lead angle pair, in theory it locks, but in practice vibration, thermal expansion, and oil-film wedge effects can drop the effective friction coefficient below the locking threshold.
The honest engineering answer: if back-drive prevention matters for safety (rigging, valve hold-open, vehicle parking), add a separate failsafe brake. Don't trust self-locking as a single-fault-tolerant feature. Theatrical rigging codes like ANSI E1.6-1 explicitly require a brake in addition to self-locking gearing.
That's run-in wear shifting the contact pattern, and it's normal up to a point. New globoid pairs ship with a slightly oversized contact patch on the wheel teeth — the bronze conforms to the steel worm during the first 20-100 hours, which initially polishes the surfaces and improves efficiency by 2-4 percentage points, not drops it.
If you're seeing a drop instead, your contact pattern is moving off the throat. Pull the cover and look at the wear marks on the wheel — they should be centred on the throat with symmetric witness marks. Marks biased to one side mean centre distance is off; marks at the tooth tips mean the worm is sitting too deep. Either way, fix it before the first 200 hours or you'll polish a permanent groove into the bronze and lose 15-20% of rated capacity.
Roughly 3,500 RPM at the worm shaft for industrial pairs, and 1,800 RPM if you want long life. Above that, the sliding velocity at the mesh exceeds the lubricant's hydrodynamic film capacity and you transition from full-film to mixed-film lubrication. Efficiency drops 10-15 percentage points and bronze wear rate climbs roughly with the cube of sliding velocity.
If you need higher input speeds, either step down with a primary helical or planetary stage first, or specify a spray-lubricated case with circulating oil rather than splash. Cone Drive's HM-series rates higher input speeds because they ship with built-in oil pumps; the cheaper splash-lubricated variants of the same gearset are speed-limited even though the gear geometry is identical.
References & Further Reading
- Wikipedia contributors. Worm drive. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
