Need custom worm gears? Contact us — WhatsApp +86 138 1403 4409 · worm@wormgearmfg.com
ISO Certified
DIN 4-5 Precision Grade
Custom Manufacturing
Design

Worm Gear Backdrive Calculations: Formula and Self-Locking

by Mike · Nov 4, 2026 · 8 min read

What Backdriving Actually Means

A gearbox is backdriven when the load, not the motor, turns the train — a raised load releasing its stored energy through the output shaft, a head of water against a valve actuator, a machine settle after power loss. With most gear types the question never comes up: spur, helical and planetary meshes roll energy backwards almost as easily as forwards. The worm gear is the exception. Its sliding mesh can act as a screw that the wheel cannot easily turn backwards, and this conditional one-way behavior — self-locking — is one of the main reasons designers choose a worm in the first place. Our self-locking worm gears guide covers the behavior in application terms; this article does the numbers: when a given mesh will hold, when it will spin back, and what to do when the calculation says it will not.

The Criterion: Lead Angle vs Friction Angle

Backdriving a worm pair means driving the worm axially by pushing on its thread — mechanically the same problem as turning a screw by pushing along its axis. The comparison that decides it is between two angles on the worm:

  • The lead angle (γ): the slope of the thread helix at the worm's pitch cylinder. It is set by the geometry — lead (starts × axial pitch) against pitch-circle circumference. More starts or a smaller pitch diameter make the helix steeper.
  • The friction angle (φ): the angle whose tangent is the coefficient of friction in the mesh, φ = arctan(µ).

The ideal criterion is then simply:

  • γ < φ — the mesh is self-locking in theory: the thread is too shallow for the wheel's push to screw the worm around; friction wins and the pair holds.
  • γ > φ — the mesh backdrives: the helix is steep enough that the axial push turns the worm, and the load spins the train backwards, usually with noticeable chattering as it accelerates.

Everything quantitative in a backdrive calculation therefore reduces to two tasks: pinning down the worm's actual lead angle, and choosing a defensible coefficient of friction.

Numbers You Need: Coefficients and Angles

Published engineering data for a hardened steel worm running on a bronze wheel under good lubrication puts the coefficient of friction in the region of 0.03 to 0.05 — with the explicit caution that lubricant, surface finish and speed all move it. Converting to friction angles:

Coefficient µ (steel worm on bronze wheel)Friction angle φ = arctan(µ)
0.03≈ 1.7°
0.04≈ 2.3°
0.05≈ 2.9°

So the friction angle a designer works against is small — under about three degrees at typical lubricated values. Lead angles, by contrast, span a wide range: single-start worms on large-ratio reductions commonly present lead angles in the low single digits of degrees, while multi-start worms on small ratios run steeper — often well above ten degrees. That spread is why the answer to "will it backdrive?" is always mesh-specific: there is no such thing as a self-locking gearbox type, only a self-locking lead-angle-to-friction-angle relationship at a given µ.

For the geometry side, the lead angle follows from the worm definition: tan(γ) = lead ÷ (π × pitch diameter of the worm). Our gear calculator page handles the arithmetic when you have module, starts and pitch diameter in hand.

Worked Examples: One Mesh That Backdrives, One That Holds

Example 1 — a 4-start worm at a 10:1 ratio. The multi-start thread gives a steep helix; taking a representative lead angle of about 15° for this class of geometry, and the most favorable lubricated friction value (µ = 0.05, φ ≈ 2.9°): 15° > 2.9°, decisively. This mesh backdrives freely, and the designer must plan for the load coming home on power loss — a brake on the input, or a backstop on the output. No realistic coefficient of friction rescues it; the margin is more than 5:1 in tangent terms.

Example 2 — a single-start worm at a high ratio. With one start on the same pitch diameter, the helix flattens; take a lead angle of about 2° for this class. Against the same friction angle of 2.9° (µ = 0.05): 2° < 2.9°, the pair holds in theory. But narrow the margin in your head: at µ = 0.03 the friction angle is only 1.7°, and 2° > 1.7° — the same physical mesh, now on better lubrication, backdrives. A mesh this close to the line should be treated as not reliably self-locking; the coefficient moved less than the margin.

The lesson from pairing the examples: high-ratio single-start worms sit near the boundary, and the boundary is made of a coefficient nobody can hold constant. If the application must not run backwards, the criterion tells you whether holding is *plausible* — never whether it is *guaranteed*.

Why the Textbook Criterion Is Not a Guarantee

The γ-vs-φ test is a static, ideal-condition comparison, and at least four real-world effects erode it:

  • Vibration. Any superimposed vibration or shock unloads the mesh momentarily; friction that is intermittently relieved does not hold a screw. A worm set that holds on the bench will creep on a vibrating conveyor frame. This is also why the industry's plain position, worth quoting in any internal review, is that worm self-locking must not be relied on as a safety device.
  • Running-in and surface changes. µ is not a material constant; it drifts with wear, lubricant condition and temperature. A margin that existed on day one can disappear in service.
  • Speed of the release. A slow, steady load release is the kindest case; an impact that moves the worm before static friction fully develops finds less resistance.
  • Everything else in the train. Bearings, seals and couplings add their own drag — sometimes helpful, sometimes the reverse when an oil seal's stiction is what is actually holding the shaft.

Treat the calculation, then, as a screening step: γ well below φ at µ = 0.03 (the optimistic end) is the only configuration that deserves the word "self-locking" at all, and even then a positive holding device is specified whenever the consequences of reverse motion are more than cosmetic.

Engineering Around Backdrive: Brakes and Backstops

When the calculation says the mesh will run backwards — or the margin is too thin to trust — the standard answers are external and positive. A brake on the high-speed input shaft multiplies its effect through the ratio and holds with margin. A backstop (a roller or sprag clutch) on the output or intermediate shaft permits drive in one direction only and is the conventional solution on conveyors and gates, where reverse rotation must be blocked absolutely rather than approximately. Hand-operated equipment adds its own variants — ratchets and worm-style winch gearing — which our custom worm gear page covers from the manufacturing side: backstop bosses, one-way lugs and the shaft details that make them practical to machine. For hand-crank driven worm gearboxes specifically, the torque and safety arithmetic belongs to the same framework used here.

Frequently Asked Questions

What is the formula for worm gear self-locking?

The screening criterion is lead angle below friction angle: γ < φ, where φ = arctan(µ). The lead angle comes from the worm geometry, tan(γ) = lead ÷ (π × worm pitch diameter), and µ for a lubricated hardened-steel-on-bronze pair is on the order of 0.03–0.05 — a friction angle under about 3°. If the lead angle is not comfortably below that, treat the mesh as backdrivable.

Can a worm gearbox be backdriven?

Many can. Multi-start, low-ratio worms have steep lead angles and backdrive freely; only high-ratio, single-start meshes fall near or below the friction-angle line, and even those can lose the margin to better lubrication (lower µ), vibration, or running-in. The practical rule: calculate the margin, then assume the load will run backwards unless a brake or backstop is fitted.

Is a self-locking worm gear safe for holding a load?

Not as the primary safety device. Self-locking is a friction condition, not a positive lock — vibration, shock and friction drift can all release it — and industry guidance treats it accordingly. Use the worm's holding tendency as a bonus, and fit a brake or backstop wherever reverse motion would create a hazard.

Mike — Sales Engineer, WormGear Mfg. Mike handles custom worm gear RFQs and technical consultation for overseas OEM buyers. He writes about worm gear manufacturing, DIN/AGMA/JIS standards, and sourcing from China.