A saw-tooth worm gear is a worm-and-wheel pair where the worm thread has an asymmetric, saw-tooth profile — one flank steep, the other shallow — so the wheel sees a different pressure angle in drive versus reverse. Typical industrial sets run 5 to 60 RPM at the wheel with holding loads up to several tonnes. We use this geometry when one direction carries the working load and the other only resets, which is exactly the case on screw hoists, jacks, and load-holding clamps like the Columbus McKinnon Lodestar chain hoist family.
Saw-tooth Worm Gear Interactive Calculator
Vary lead angle, friction, and asymmetric flank angles to see drive efficiency, back-drive tendency, and self-locking margin.
Equation Used
The worked example diagram identifies a 14.5 deg working flank and a 35 deg back flank. This calculator applies those asymmetric flank angles to the standard worm-drive friction relation: forward drive uses the working flank, while back-drive locking is checked against the larger back-flank friction angle.
- Lubricated steel worm on bronze wheel.
- Working flank carries forward drive; back flank controls back-drive locking.
- Pressure angle effect is included as mu / cos(alpha).
- Positive holding margin means the back flank is self-locking.
How the Saw-tooth Worm Gear Works
The trick of a saw-tooth worm is that the two flanks of each thread do different jobs. The steep working flank — typically a pressure angle near 14.5° — carries the load with high contact stress, generous oil film thickness, and low sliding loss. The shallow back flank sits at something closer to 30° to 40° and behaves like a wedge: when the wheel tries to back-drive the worm, that flank converts torque into a normal force the worm bearings simply can't overcome at any practical efficiency. Result — the gear self-locks in one direction without needing a separate brake.
The geometry only works if you respect the lead angle. Below roughly 5° the pair is reliably self-locking but efficiency drops below 40%, so the gearmotor has to be sized large. Push the lead angle past 10° and you start losing the holding behaviour — under vibration the wheel will creep, which is the failure you'll see on a worn screw jack that won't hold position overnight. The single-flank load profile means the worm wears asymmetrically, and if you flip the input direction by mistake (say a wired-backwards 3-phase motor), the shallow flank takes a load it was never cut for and pits within a few hundred cycles.
Material pairing matters here more than on a symmetric worm. We run a hardened and ground steel worm — 58 HRC minimum on the working flank — against a centrifugally cast phosphor bronze wheel. The bronze deliberately wears in to match the worm profile during the first 20 to 50 hours of running, and the asymmetric profile means the wear-in pattern is visibly lopsided. That is normal. What is not normal is scoring on the back flank — that tells you the drive has been back-driven hard, usually a sign of a missing or failed input brake on a hoist application.
Key Components
- Saw-tooth Worm (input): The hardened steel screw with the asymmetric thread profile. Working flank near 14.5° carries the drive load; back flank at 30°-40° provides the self-locking wedge. Surface finish must be Ra 0.4 µm or better on the working flank to keep contact stress under 1200 MPa at rated load.
- Worm Wheel (output): Centrifugally cast phosphor bronze (CuSn12 typical) with teeth cut to match the asymmetric worm. Tooth count usually 30 to 60, giving single-stage ratios of 30:1 to 60:1. The bronze conforms to the worm during run-in, which is why you cannot mix worms and wheels between gearboxes.
- Worm Bearings: Tapered roller pair on the worm shaft, taking the axial thrust generated by the wedge action of the back flank. On a 5 kN holding load the axial force at the worm can hit 3 to 4 kN, so undersized angular contact bearings are a common failure point.
- Centre Distance Housing: Cast iron or aluminium housing holding worm and wheel at fixed centre distance, typically held to ±0.05 mm. Too tight and the pair runs hot and seizes during run-in; too loose and backlash exceeds the 0.1° to 0.3° spec, which on a hoist shows as a perceptible drop when the load is released onto the gear.
- Oil Bath Lubrication: ISO VG 460 to VG 680 mineral or PAO gear oil, filled to the worm centreline. The high sliding velocity at the worm-wheel contact — often 3 to 8 m/s — generates heat that must be dumped through the housing, so any reduction in oil level pushes contact temperature past 90 °C and accelerates bronze wear.
Real-World Applications of the Saw-tooth Worm Gear
Saw-tooth worm gears earn their place anywhere the load only needs to be driven one way and held the other. That covers a lot of real industrial hardware — hoists, screw jacks, valve actuators, slew drives, and clamp mechanisms. The common thread is a duty cycle dominated by holding rather than running, where the self-locking back flank replaces what would otherwise be a mechanical or electrical brake.
- Material handling: Lifting screw drive on a Columbus McKinnon Lodestar L-series electric chain hoist — saw-tooth worm holds the load when power drops, eliminating the need for a separate load brake on lower-duty units.
- Stage and theatre rigging: Counterweight-assist winches on the JR Clancy PowerLift system at venues like the Sydney Opera House, where the self-locking worm guarantees the batten cannot creep down between cues.
- Solar tracking: Slew drive on a Kinematics H-series solar tracker — the asymmetric profile holds 200 kN-m of wind reaction torque without the inverter having to apply continuous holding current.
- Industrial valves: Manual override gearbox on a Rotork IW-range quarter-turn actuator on a 12-inch butterfly valve in a refinery, where the saw-tooth worm prevents the valve drifting open under fluid pressure.
- Heavy machine tool: Tailstock clamping screw on a Schiess vertical turret lathe — clamp force of 80 kN must hold all day with the input shaft completely de-energised.
- Mining: Brake-release and tensioning drive on a Joy Global longwall conveyor head end, where the worm holds chain tension during a stoppage so the drive doesn't unwind under load.
The Formula Behind the Saw-tooth Worm Gear
The number that decides everything on a saw-tooth worm is the drive-direction efficiency, because it sets the input torque the motor has to deliver and tells you whether the pair will self-lock. At the low end of the typical lead angle range, around 3°, you get bombproof self-locking but efficiency drops near 30% — the motor has to be sized 3× the output torque demand. The sweet spot sits at 5° to 7° lead angle, where efficiency lands in the 45% to 55% band and self-locking is still reliable. Push past 10° lead and efficiency climbs above 70%, but you lose the holding feature entirely and the wheel will back-drive under vibration.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| ηdrive | Forward-drive efficiency of the worm-wheel pair | dimensionless (0-1) | dimensionless (0-1) |
| λ | Lead angle of the worm thread (working flank) | degrees | degrees |
| φ | Friction angle, where tan(φ) = μ, the sliding friction coefficient at the worm-wheel contact | degrees | degrees |
| μ | Sliding friction coefficient between hardened steel worm and bronze wheel under oil bath lubrication | dimensionless (typ. 0.03-0.06) | dimensionless (typ. 0.03-0.06) |
Worked Example: Saw-tooth Worm Gear in a stage rigging counterweight winch
You are sizing the saw-tooth worm-and-wheel reduction on a JR Clancy-style PowerLift batten winch holding a 4500 N lighting truss at a regional repertory theatre. The wheel must turn at 20 RPM during cue moves, and the load must hold solid between cues with the motor de-energised. Worm is single-start, working-flank lead angle 6°, back-flank wedge angle 35°. Lubrication is ISO VG 460, giving μ ≈ 0.04 at running speed. Wheel has 40 teeth.
Given
- λ = 6 degrees
- μ = 0.04 dimensionless
- Output load = 4500 N
- Wheel teeth Zw = 40 teeth
- Wheel speed = 20 RPM
Solution
Step 1 — convert friction coefficient to friction angle:
Step 2 — compute drive efficiency at the nominal 6° lead angle:
That 72% is the headline number — for every kW the motor delivers, 720 W reaches the wheel and 280 W becomes heat in the oil. On a 4500 N truss at 20 RPM the wheel work rate is modest, but on a 50 kN industrial hoist that 28% loss is what dictates oil cooler sizing.
Step 3 — check the low end of the typical lead-angle range, 3°, which is where self-locking is most aggressive:
At 3° lead the same motor only delivers 57% to the load — you would need to upsize the gearmotor by roughly 25% to lift the same truss at the same speed. The trade is rock-solid holding even with worn bronze.
Step 4 — check the high end, 10° lead angle, where designers push to recover efficiency:
81% efficiency looks attractive but the self-locking margin collapses. The static reverse condition wants λ < φstatic, and with static μ around 0.10 (φ ≈ 5.7°) the 10° pair will absolutely back-drive under a vibrating truss. You would need to add a separate load brake — defeating the point of choosing a saw-tooth worm in the first place.
Result
Forward-drive efficiency at the nominal 6° lead angle is 72%, so the motor sees roughly 4500 N / 0. 72 ≈ 6250 N of equivalent output demand at the lifting screw, and the gearbox dumps about 28% of input power as heat in the oil bath. Across the typical lead-angle range the answer slides from 57% at 3° (heavy motor, bombproof holding) through 72% at 6° (the sweet spot for theatre and light industrial hoists) to 81% at 10° (efficient but no longer self-locking) — the 6° design is exactly where most stage and small-hoist builders land for that reason. If your measured efficiency comes in 10 to 15 points below predicted, the usual suspects are: (1) oil level below the worm centreline starving the contact, (2) worm surface finish above Ra 0.8 µm from a re-ground worm not properly lapped, or (3) tapered roller bearings on the worm shaft over-preloaded during reassembly, which adds parasitic torque that scales linearly with input speed.
Saw-tooth Worm Gear vs Alternatives
Saw-tooth worm gears compete against symmetric worm pairs and against helical/bevel reducers paired with separate brakes. The decision usually comes down to whether you actually need the asymmetric holding behaviour, and whether you can tolerate the efficiency penalty that comes with any worm.
| Property | Saw-tooth Worm Gear | Symmetric Worm Gear | Helical Reducer + Brake |
|---|---|---|---|
| Forward-drive efficiency at typical lead angle | 55-75% (asymmetric flank loss) | 60-90% (depending on lead) | 92-97% |
| Self-locking under static load | Yes — by design, single-direction | Conditional — only λ < ~5° | No — requires separate brake |
| Output speed range | 5-60 RPM typical | 5-300 RPM | 10-3000 RPM |
| Holding load capacity | Up to several tonnes static | Up to several tonnes static | Limited by brake torque rating |
| Cost per kW of rated output | Mid — bronze wheel + ground worm | Low — standard tooling | High — gearbox + brake assembly |
| Service life under reversing load | Poor — back flank pits if reversed under load | Good — symmetric wear | Excellent — designed for reversing |
| Backlash at the wheel | 0.1-0.3° (asymmetric tooth fit) | 0.05-0.2° | 0.02-0.1° with anti-backlash |
| Best application fit | One-way lifting and clamping | General reduction with light holding | High-speed reversing drives |
Frequently Asked Questions About Saw-tooth Worm Gear
Self-locking is a static-friction phenomenon, and the friction coefficient drops sharply when the contact is vibrated. On a hoist sharing a building with HVAC plant, compressors, or a fly tower with running motors, the worm-wheel contact is being micro-vibrated continuously, which moves the effective friction angle from around 5.7° (static) down toward 2.3° (dynamic).
If your worm lead angle is at the high end of the design range — say 8° to 10° — you cross from self-locking into back-drivable as soon as ambient vibration is present. The fix is either a lower-lead-angle worm or, more commonly, a small fail-safe disc brake on the worm shaft. Don't try to compensate with thicker oil; that just hides the symptom for a few weeks.
Saw-tooth worm-and-wheel pairs run-in together. During the first 20-50 hours the bronze wheel teeth deform plastically to match the asymmetric worm profile, and the wear pattern is direction-specific — the working flank wears in deeply, the back flank barely at all.
When you put a new worm into a wheel that's already worn-in to a different worm, contact area collapses to maybe 15-20% of the design footprint. Hertzian stress on those small patches blows past 1500 MPa, you get noise (squeal under load), heat (oil temperature climbs above 90 °C), and accelerated bronze wear. Standard practice is to replace worm and wheel as a matched set, always.
Run the duty-cycle calculation. If holding time is more than about 95% of total cycle time, choose the 3° design — efficiency doesn't matter much because the motor barely runs, and the steeper safety margin on self-locking lets you skip a brake entirely.
If running time approaches 20-30% of cycle (e.g. a positioning jack on an automated tool changer), the 6° design pays back in motor sizing — you can use a smaller gearmotor and a smaller VFD because the pair delivers 72% rather than 57%. Below 6° you also tend to hit wheel-tooth bending stress limits earlier because the contact ratio is lower, so 3° pairs need a slightly oversized wheel module.
This usually points to a load-reversal event you didn't see. Common culprits: (1) a coupling slip on shutdown that let the load swing the wheel briefly backwards, (2) a control fault that reversed the motor for a fraction of a second during a brake test, or (3) a thermal shutdown where the load fell against the back flank as the brake (if any) released slowly.
The back flank's 30-40° geometry was designed to wedge under load, not to roll under load — Hertzian contact stress on that flank under reverse driving is roughly 2× the working flank, so even a few seconds of reverse drive at full load is enough to initiate pitting. Inspect your control sequence and any electrical interlocks before swapping the worm.
Marginally, and only at the cost of efficiency. Thicker oil (say going from VG 460 to VG 680) raises the boundary-film friction coefficient from around 0.04 toward 0.05-0.06, which moves the friction angle and gives you a tiny extra self-locking margin under vibration.
What you'll actually notice is a 5-10% drop in forward-drive efficiency and a 10-15 °C rise in operating oil temperature, both of which shorten the bronze wheel's service life. The right answer when holding margin is inadequate is to change the worm lead angle or add a brake — not to fight the problem in the lubricant.
No — and that is the design call most often made wrong. If the duty cycle is dominated by running, you are paying a 5-15 percentage-point efficiency penalty (versus a symmetric worm at the same lead angle) for a holding feature you barely use. That shows up as higher motor power draw, larger oil cooler, and shorter bronze life.
Above roughly 30% running duty, switch to either a symmetric worm pair sized for your lead angle, or a helical reducer with a fail-safe brake. Saw-tooth geometry only earns its keep when holding load dominates the duty cycle.
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.
