Tout savoir sur les engrenages à vis sans fin et les avantages de collaborer avec un fabricant

What Is a Worm Gear? (Définition)

UN vis sans fin — also called a worm drive ou boîte de vitesses à vis sans fin when housed as a reducer — is a gear system that transmits motion and power between two non-parallel, non-intersecting arbres, almost always at a 90° right angle. It has two components:

• Le ver — a shaft with a continuous helical thread, essentially a screw. It is the input member and usually the driving element.
• The worm wheel (vis sans fin) — a toothed wheel whose curved teeth are conjugate to the worm’s helical profile. It is the output member. (Note: some sources use “vis sans fin” for the wheel alone; in this article vis sans fin means the complete assembly.)

Because the shafts are perpendicular and offset, a worm gear set always changes the direction of rotation by 90° — a geometric property no other en une seule étape reducer provides in such a compact envelope. The idea is ancient: Archimedes described the screw mechanism around 250 BC, et le “Archimedean worm” is still a basic form today.

Three numbers define any worm gearbox at the most fundamental level: le ratio de réduction (how much the output slows), le couple de sortie (continuous force the output can deliver), et le center distance (distance between the two shaft axes, which sets the physical size). Everything else on a spec sheet — thermal rating, facteur de service, contrecoup, mounting — flows from these.

Why Engineers Choose Worm Gears

• High single-stage ratio: 5:1 pour 100:1 in one stage, where spur/helical gears need two or three stages to reach the same ratio.
• Self-locking option: at high ratios the output mechanically locks when the motor stops — no brake needed.
• Calme, mouvement fluide: teeth engage gradually over a large contact area; typical 55–62 dB vs 62–75 dB for helical units.
• Compact right-angle package: small footprint for the torque delivered.
• Shock-load tolerance: the conforming (line/area) contact spreads load well.

The trade-off is efficiency and heat. The mesh is mostly sliding contact, so a large share of transmitted power becomes heat — the single most misunderstood and misspecified parameter in worm-gear selection.

How a Worm Gear Works: Étape par étape

The worm-drive action is best understood as a screw turning a wheel — but the physics is helical-gear meshing, not pure threading.

1. The worm rotates as the input. Its helical thread is the driving surface. Rotating it one full turn advances the worm wheel by exactly the number of thread starts.
2. The thread pushes the wheel teeth. As the worm turns, its flank contacts a wheel tooth, slides along it, engages fully, and releases — a continuous, progressive meshing. The helix direction and pressure angle of the worm must match the wheel, and the worm’s axial pitch must equal the wheel’s circular pitch, or the set will not mesh.
3. Speed is reduced by the tooth count. A single-start worm advances the wheel one tooth per revolution; a 40-tooth wheel therefore gives 40:1. Two starts give 20:1 from the same wheel. The ratio is set purely by geometry: i = Z₂ / Z₁ (wheel teeth ÷ worm starts).
4. Torque is multiplied. By conservation of power (ignoring losses), the slower output carries proportionally more torque: T_out =T_dans × je × n. This is why a small worm can hold a heavy load.
5. Sliding dominates the contact. Unlike spur/helical gears that roll, the worm thread slides across the wheel face. That sliding is what creates high friction, efficacité inférieure, chaleur, and — at a shallow lead angle — self-locking.
6. Self-locking appears when the angle is shallow enough. If the worm’s lead angle (γ) is smaller than the mesh friction angle (ρ′), the wheel cannot push the worm backward; the output locks when the motor de-energizes.

Single vs. Multi-Start Worms

The number of starts (thread heads, Z₁) is the primary lever for both ratio and efficiency:

• Single-start (Z₁ = 1): shallow lead angle, rapport élevé, strong self-locking, but low efficiency (40–50%).
• Double-start (Z₁ = 2): moderate lead angle, less self-locking, efficiency 50–78%.
• Multi-start (Z₁ = 3–4): steep lead angle, haute efficacité (78–92%) but generally not self-locking.

Worm Gear Types Compared

Worm sets are classified by how the worm and wheel “wrap” around each other. The enveloping form dictates contact area, load capacity, et le coût.

Enveloping vs. planar double-enveloping: a standard enveloping worm has a cylindrical worm with a throated wheel. UN double-enveloping (par ex., Cone Drive’s Envex®) throats both members so the teeth maintain uniform contact through the mesh — highest load capacity but the most demanding to machine. For most industrial NMRV-class reducers, single-enveloping is the cost/performance sweet spot.

Vis sans fin vs. Other Gear Types

Choosing worm over spur, hélicoïdal, bevel, or planetary depends on ratio, efficacité, autobloquant, and envelope.

Règle générale: pick a worm drive when you need a large right-angle reduction, quiet running, or self-locking in one stage. Pick spur/helical/planetary when efficiency and continuous-duty thermal load dominate (par ex., high-speed conveyors, traction). See also our gearbox vs. motoréducteur guide.

Worm Gear Efficiency by Ratio

Efficiency falls monotonically with ratio because a higher ratio forces a shallower lead angle. The table consolidates AGMA / ISO 14521 methodology with manufacturer acceptance-test data (run-in unit, ~40°C oil, PAO synthetic — the most favorable normal condition). Cold-start runs 5–10% lower.

Material Pairing: Hardened-Steel Worm + Bronze Wheel

The worm and wheel are deliberately different hardnesses. A hard steel worm against a softer bronze wheel prevents adhesive wear (galling): the bronze sacrifices itself slowly while protecting the more expensive worm. This is why bronze-wheel replacement is the most common major service item (typically 6–10 years of continuous duty).

Worm Gear Engineering Data: Formules & Limits

These are the equations an engineer actually uses when specifying a worm drive. All conform to AGMA 6034-B92 (ANSI/AGMA 6034) et ISO 14521 methodology.

1. Ratio de réduction

i = Z2 / Z1 — wheel teeth (Z₂) ÷ worm starts (Z₁). Keep Z₂ > 28 to avoid undercutting.

2. Lead angle

γ = arctan( Z1 · m / d1 ) = arctan( Z1 / q ) — where m is the axial module, d₁ the worm pitch diameter, and q = d₁/m the diameter factor. The lead angle is the single most important efficiency variable.

3. Efficacité (worm driving)

η = tan γ / tan( γ + ρ′ ), where the virtual friction angle ρ′ = arctan( μ / cos αn ), μ is the mesh friction coefficient (set by sliding velocity, lubricant, material, température), and α_n is the normal pressure angle (typically 20°, so cos α_n ≈ 0.940).

The practical back-drive form is also useful:

ηback = tan( γ − ρ′ ) / tan γ

4. Self-locking condition

Self-locking when γ < ρ′ (equivalently, when λ < φ, the equivalent friction angle). Below roughly 5° lead angle, most sets self-lock. Remember the trade: the same shallow angle that gives self-locking also gives the lowest efficiency (often < 50%).

5. Couple de sortie

Tout = Tin × i × η — always include η; à 50:1/55% you lose 45% of input torque to friction.

6. Center distance

a = m ( q + Z2 + 2x ) / 2 — where x is the addendum-modification coefficient. Center distance sets the unit size and is fixed by the housing; getting it wrong causes misalignment wear.

7. Thermal limit (the rating that actually bites)

The AGMA power rating is based on 10 h of continuous uniform load and is governed by pitting/wear resistance et by heat dissipation. Keep oil temperature ≤ 80–90°C; above this, lubricant viscosity collapses and the mesh risks scuffing failure. The allowable thermal power is Pth = ks · A · ( Toil − Tambient ).

Worked example — why efficiency is a thermal problem, not a footnote

Considérez un 3 kW conveyor drive at 50:1, continu 18 h/day. À 75% efficacité, lost power = 3 × 0.25 = 0.75 kW of heat, continuously. À 55% efficacité (single-start, autobloquant), lost power = 3 × 0.45 = 1.35 kW. Le 0.6 kW difference behaves like a 600 W space heater sealed inside the housing — raising oil temperature 15–20°C, dropping viscosity ~40%, and triggering a self-reinforcing scuffing failure months later. The fix is a multi-start worm (higher γ, higher η) or forced/larger cooling — not a bigger motor.

Best Applications for Worm Gears

Worm drives earn their place wherever a large right-angle reduction, quiet running, or self-locking matters more than peak efficiency.

For battery- or efficiency-critical motion where a worm’s losses are unacceptable, notre Moteurs BLDC paired with réducteurs planétaires are the usual alternative.

How to Select a Worm Gear: Étape par étape

1. Define the ratio. From required output speed: je = n_dans / n_out. Worm sets cover 5:1–100:1 in one stage.
2. Define the output torque. Load torque × service factor. Then back-calculate the needed input torque: T_dans =T_out / ( je × n ).
3. Pick the start count (Z₁). Need self-locking? Use single-start. Need efficiency for continuous duty? Use double/multi-start. This is the key efficiency lever.
4. Size the center distance. Choose the smallest standard center distance whose AGMA-rated thermal and pitting capacity exceeds your continuous power at the operating ratio. Ne pas size on peak torque alone.
5. Select the material pair. Hardened-steel worm + phosphor-bronze wheel for general duty; aluminum bronze for heavy/reversing; cast iron only for light duty.
6. Specify lubrication & refroidissement. ISO VG220 PAO synthetic as standard; VG320 for high-cycle reversing; PAG for high temperature. For continuous high-load, confirm the housing can shed the lost heat (P_th).
7. Confirm self-locking (if required). Verify γ < ρ′ at your operating temperature — note that hot lubricant lowers μ and can break self-locking on borderline designs.
8. Verify mounting & alignment. Match the motor flange (IEC B5/B14 or NEMA C-face), then ensure coaxiality, perpendicularity, and the correct center distance at installation to avoid eccentric wear.

Common Worm Gear Engineering Mistakes

• Sizing on peak torque, ignoring thermal rating. The AGMA rating is a 10-hour continuous-duty limit governed by heat. A unit fine for a 2-minute peak will cook under 18-hour continuous load. Always check P_th.
• Assuming high efficiency. À 50:1 a self-locking worm is ~55% efficient; you lose ~45% of input torque and ~1.3 kW of 3 kW as heat. Budget motor power and cooling accordingly.
• Assuming it can be back-driven. Self-locking sets (single-start, rapport élevé) cannot be reversed by the load. Using one where the application needs back-driving motion is a design error — and relying on self-locking as the sole safety brake can fail if the oil warms up.
• Using the wrong lubricant. Worm meshes need EP worm-grade oil (ISO VG220/320), not ordinary engine or gear oil. The wrong oil raises μ, drops efficiency, and causes scuffing.
• Wrong alignment / center distance. Off-center or non-perpendicular mounting creates eccentric load and poor mesh, accelerating wear and noise.
• Choosing the wrong start count. Specifying a single-start when efficiency matters (continuous conveyor) or a multi-start when self-locking is safety-critical. Decide Z₁ from the efficiency/self-lock trade first.

Worm Gear Troubleshooting: Problème → Cause → Solution