The two formulas every worm drive comes back to

Forget the long lists of pitch and module equations for a moment. Two formulas determine 90 percent of the design decisions on a worm and worm wheel pair, and most calculation mistakes in the field come from misapplying these two — not from advanced geometry.

The reason these two formulas matter so much is that they capture the central trade-off of worm gearing — high ratio means low efficiency, low ratio means high efficiency, and you cannot have both in the same set. The second formula explains the first one’s hidden cost.

Reading the ratio formula correctly

The integer trap that ruins clean textbook ratios

Both Z₁ and Z₂ must be integers, and that constraint matters more than most calculators acknowledge. If a customer asks for “exactly 35:1,” our engineering desk has to tell them they will get one of three nearest practical ratios: Z₂ = 35 with Z₁ = 1 gives 35:1 exactly, Z₂ = 70 with Z₁ = 2 gives 35:1 exactly, or Z₂ = 36 with Z₁ = 1 gives 36:1 (a 2.9 percent overshoot). The choice depends on what else the application needs — Z₂ = 35 is fine for a single-start drive, Z₂ = 70 doubles the wheel diameter at the same module, and Z₂ = 36 is a small compromise that lets you use a more common wheel size.

Asking for 35.5:1 simply does not work — there is no integer pair that delivers exactly that ratio. The clean number on the design sheet has to round to something the factory can actually cut. For Korean and Japanese OEM applications where downstream encoders and motor controls assume an exact gear ratio, this rounding has to happen at the design stage, not after the parts are made.

How efficiency follows from lead angle

The lead angle λ is the angle between the worm thread helix and a plane perpendicular to the worm axis. For a single-start worm with small pitch diameter, λ might be 3 to 5 degrees. For a 4-start worm with the same pitch diameter and module, λ climbs to 15 to 20 degrees. The relationship is geometric: more starts at the same module means a steeper helix.

Plug numbers into the efficiency formula and the trade-off becomes concrete. Assume a friction angle φ = 6 degrees, which is realistic for well-lubricated steel-on-phosphor-bronze:

The shape of the curve matters. Going from 3° to 10° lead angle nearly doubles efficiency. Going from 20° to 30° barely moves the needle. The sweet spot for high-efficiency multi-start drives sits around 15 to 20 degrees lead angle — beyond that you get diminishing returns and you start losing the wheel face width that makes the drive serviceable. Most catalogue worm and worm wheel pairs sit in two clusters: 3 to 5 degrees (high-ratio self-locking) or 12 to 18 degrees (medium-ratio efficiency-driven).

Worked example 1 — Conveyor belt drive

Notice the rounding decision at step 5 — the exact arithmetic ratio is 29.4:1, but the nearest practical integer-tooth ratio is 30:1, which gives a slightly slower belt speed. The customer accepted that compromise without a visible difference at the conveyor output. This is normal for industrial drives.

Worked example 2 — Hoist drum drive

A small workshop hoist lifts up to 500 kg on a 100 mm radius drum. Lifting speed is specified at 6 m/min. Self-locking is mandatory because a falling load would be a safety hazard. The customer wants to use a standard 1,400 rpm three-phase motor.

Self-locking eliminates multi-start worms — we are forced into a 1-start design with a low lead angle, accepting the efficiency penalty.

Two things stand out from this calculation. First, the efficiency penalty for self-locking is substantial — about 67 percent of input power becomes heat in the drive. Second, the motor power requirement (3 kW) is much higher than the same load would need in a high-efficiency helical right-angle reducer (perhaps 1.5 kW). The customer is paying for self-locking with extra electricity over the lifetime of the hoist. For a workshop hoist running maybe 200 hours per year, that trade-off is acceptable. For a 24-hour production hoist, it would not be — the right answer there is helical reducer plus a separate mechanical brake.

Worked example 3 — Indexing rotary table

A 4-station indexing table positions automotive seat-frame welding fixtures. Each station holds 12 kg, total table mass 80 kg, table radius 400 mm. Index time per station is 1.2 seconds (90 degrees rotation). Holding torque between motions must resist accidental nudge but the drive itself is electrically held by a servo brake — self-locking is desirable but not mandatory.

The takeaway: dynamic indexing applications drive the calculation through acceleration torque, not steady-state torque. The polar moment of inertia of the table itself often dominates over the workpiece mass, especially on heavy steel rotary tables. Servo selection has to satisfy peak torque, not mean torque — failing this is the most common reason indexing prototypes stall during the first cycle.

Common calculation mistakes that destroy designs

Confusing Z₁ and Z₂. A surprising number of first drawings arrive with worm starts and wheel teeth swapped — somebody wrote the wheel tooth count where the formula expects the worm starts. The result is a calculated ratio of 1/40 instead of 40, which makes the math look absurd and stops the design dead. Always label clearly: Z₁ for the worm, Z₂ for the wheel.

Forgetting to divide by efficiency. The basic ratio gives you the kinematic relationship between input and output speeds. Translating that to torque requires dividing by efficiency. Skip the efficiency divisor and you specify a motor that is far too small. The drive will stall under nominal load. Input torque = output torque ÷ (ratio × efficiency), always.

Treating efficiency as a constant. The published rated efficiency is at rated load. Light-load efficiency is much lower because the friction torque inside the gearbox stays roughly constant while useful torque shrinks. Always use the operating-point efficiency, not the headline number.

Using static torque for dynamic applications. Indexing tables, hoists with shock loads, and any drive with frequent start-stop cycles must be sized for peak acceleration torque, not steady-state torque. The peak can be 2 to 4 times the steady value depending on cycle time.

Demanding non-integer ratios. Asking for 47.3:1 has no solution. Round to the nearest practical integer-tooth ratio at the design stage. If the controller downstream needs an exact ratio, design the gear ratio first and let the controller scaling adjust to the actual ratio.

Forgetting service factor. A drive sized exactly to the calculated nominal torque has zero margin for line voltage variation, ageing, occasional overload, or thermal cycling. Apply a service factor between 1.3 (light intermittent) and 2.5 (heavy shock-loaded) before selecting the motor and gear set.

Frequently asked questions

The arithmetic on a worm and worm wheel pair is straightforward but unforgiving. Get the basic ratio formula wrong and the math nonsense is immediately visible. Get the efficiency calculation wrong and the drive ships, runs hot, fails warranty, and the mistake hides for months until the field returns start coming back. The two formulas at the start of this article carry essentially the entire load — they just have to be applied at the actual operating point, with realistic friction estimates, and rounded to integer tooth counts that the factory can actually produce.

For Korean and Japanese OEM design teams that want a calculation reviewed before committing to motor and ratio specifications, our engineering desk runs a worm gear ratio calculation review against your duty cycle, applies realistic efficiency at the actual operating point, and recommends a tooth pair that the factory can deliver to standard catalogue lead time. Standard catalogue ratios from 5:1 to 100:1 are stocked across our single-start and multi-start worm gear sets for modules M1 through M8, and custom ratios outside the catalogue range are made to order against drawing.

Editor: Cxm