Run a fingernail across the worm thread — you can feel the difference between Ra 1.6 hobbed and Ra 0.4 ground. Surface finish is the friction language of every worm gear pair, and one process step can double service life.
Worm gear surface finish is measured as average roughness Ra (micrometres) on the tooth flank — typically 1.6 to 3.2 µm for hobbed-only worm threads, 0.4 to 0.8 µm for ground threads, 0.2 to 0.4 µm for lapped pairs, and 0.1 to 0.2 µm for polished or superfinished surfaces destined for sanitary or precision applications. Each step down in Ra adds approximately 30 to 50 percent service life through improved elastohydrodynamic film formation and reduced micro-pitting. Cost premium roughly doubles each step: hobbed 1.0×, ground 1.5×, lapped 2.0×, polished 2.5×. The right finish for a given worm gear pair depends on the application duty class, lubrication regime, and required service life — not on a generic “smoother is better” rule. Most industrial worm gear pairs operate well at Ra 0.4 to 0.8 µm; precision indexers and high-power applications justify the lower roughness; food and pharmaceutical applications mandate Ra ≤ 0.4 µm regardless of mechanical need.
Worm gear contact is sliding contact. The worm thread does not roll across the wheel tooth flank in the way two spur gear teeth roll against each other; it slides, with the contact line tracing across the flank as the worm rotates. Sliding contact means friction; friction means heat, wear, and energy loss. The surface finish of the contacting flanks is the primary control variable for all three.
When a worm gear pair operates under load, the lubricant separates the two surfaces by a thin film — typically 0.3 to 1.5 micrometres thick at the contact zone. The ratio of film thickness to surface roughness is called the lambda ratio, and it determines the lubrication regime. Lambda greater than 3 means full elastohydrodynamic separation — the surfaces never touch, wear is governed by oxidation rates rather than mechanical contact. Lambda between 1 and 3 is mixed lubrication — partial contact between high spots, moderate wear. Lambda less than 1 is boundary lubrication — extensive metal-on-metal contact, accelerated wear, risk of scuffing.
The film thickness is set by oil viscosity, sliding velocity, and contact pressure — not directly by surface finish. But the surface finish sets the denominator of the lambda ratio. A worm gear pair with film thickness 0.6 µm and surface roughness Ra 0.8 µm has lambda 0.75 (boundary regime). The same pair finished at Ra 0.2 µm has lambda 3.0 (full EHL regime). Same operating conditions, same lubricant, same load — radically different lubrication and wear behaviour, set entirely by the surface finish of the worm thread and wheel teeth.
Four manufacturing processes dominate worm gear flank finishing — hobbing alone, hobbing followed by grinding, hobbing plus grinding plus lapping, and full polishing or superfinishing. Each process steps down the achievable Ra by roughly half and approximately doubles process cost.
The choice between processes is rarely about absolute Ra target; it is about the lambda ratio the application needs and the duty class the worm gear pair will see in service.
| Proces | Ra (µm) | Kostenratio | Life impact | Sollicitatie |
|---|---|---|---|---|
| Hobbed only | 1.6 – 3.2 | 1,0× | Baseline | Economical industrial |
| Hobbed + ground | 0.4 – 0.8 | 1.5× | +30–50% | Standard precision |
| Ground + lapped | 0.2 – 0.4 | 2.0× | +50–100% | High-precision indexers |
| Polished / superfinished | 0.1 – 0.2 | 2.5× | +80–150% | Sanitary, premium |
Hobbed only. The simplest finishing path. The worm thread is cut on a thread-grinding machine or generated by hobbing, and the surface as it comes off the cutter is the final surface. Achievable Ra is 1.6 to 3.2 µm depending on cutter sharpness and feed rate. Adequate for low-load, low-speed industrial worm gear pairs operating at lambda greater than 1.
Hobbed plus ground. After hobbing or thread grinding, the worm thread is finish-ground on a precision thread grinder using a vitrified or resin-bonded grinding wheel. Achievable Ra is 0.4 to 0.8 µm. This is the standard finish for industrial worm gear pairs intended for moderate to heavy continuous duty. The standard ZI involute and ZK cone-ground profiles both fall in this category.
Ground plus lapped. After grinding, the worm and wheel are lapped together as a matched pair using fine abrasive paste. Lapping removes the highest peaks of the ground surface and produces a mirror-smooth finish at Ra 0.2 to 0.4 µm. The process also self-corrects small contact pattern errors because the lapping action concentrates at the high spots. Lapped worm gear pairs are typically delivered as matched sets that cannot be substituted individually.
Polished or superfinished. Various processes produce Ra below 0.2 µm: chemically-assisted vibratory polishing (sometimes called isotropic superfinishing or REM finish), polish-grinding with a soft-bond wheel, or hand-lapping with very fine compound. Achievable Ra is 0.1 to 0.2 µm. Cost premium is significant; the process is reserved for sanitary applications where regulation mandates Ra ≤ 0.4 µm, premium high-power applications where every percentage of efficiency improvement justifies the cost, and very-low-noise applications where the smoothness improves NVH performance.
A common observation that confuses first-time worm gear specifiers: the bronze wheel surface roughness measured 6 months into service is significantly smoother than the as-manufactured surface. A wheel finished at Ra 1.6 µm during hobbing typically reads Ra 0.6 to 0.8 µm after run-in. The smoothing is real and beneficial — soft bronze running against hard steel preferentially wears the bronze peaks down to a polished surface that matches the harder steel worm thread profile. The effect is part of the natural run-in process, not a defect. Specifying Ra 0.4 µm on the bronze wheel as-manufactured is therefore over-specified for many industrial applications because run-in achieves Ra 0.6 to 0.8 µm naturally within the first 100 to 300 operating hours. The cost saving from accepting Ra 1.6 µm as-manufactured plus a defined run-in protocol can be 200 to 400 USD per worm gear pair compared to specifying lapped finish from new. The exception is precision applications where the dimensional change during run-in itself is unacceptable — for those, lapping is required to stabilise geometry from day one.
A worm gear pair has two parts and two different surface finish considerations. The hard steel worm and the soft bronze wheel face very different in-service conditions, and their finish specifications follow different rules.
The worm — hardened steel, the finish does not change. Case-hardened or through-hardened steel worms (typically 16MnCr5 case-hardened to HRC 58-62 or 42CrMo4 through-hardened to HRC 30-40) maintain their as-manufactured surface finish throughout service life. The harder material does not wear significantly under the contact stresses produced by the bronze wheel. Whatever finish the worm leaves the factory with is essentially the finish it has 10 years later. Surface finish on the worm therefore needs to be correct from day one.
The wheel — soft bronze, the finish improves with use. Phosphor bronze, aluminium bronze, or cast iron wheels start with a hobbed or shaved finish that is typically Ra 1.6 to 3.2 µm. During the first 100 to 300 hours of operation, the soft bronze peaks wear preferentially and the surface smooths to roughly Ra 0.4 to 0.8 µm — the run-in finish. The wheel surface finish is therefore self-correcting up to a point. Beyond that point, however, continued sliding contact removes material progressively and the wheel loses its tooth flank shape; this is the wear failure mode covered separately.
Implication for specification. The worm finish should be specified at the operating regime target — Ra 0.4 µm for full EHL operation, Ra 0.8 µm for mixed lubrication, Ra 1.6 µm only for low-load economical applications. The wheel finish should be specified at one tier rougher than the worm (e.g., Ra 0.8 if worm is Ra 0.4) because the wheel will run in to match the worm anyway. Over-specifying the wheel finish wastes money — under-specifying the worm finish creates a permanent operational problem.
Elastohydrodynamic film thickness in worm gear contact depends on entrainment velocity, dynamic oil viscosity, and contact pressure. The Dowson-Higginson formula gives film thickness h₀ proportional to viscosity to the 0.7 power and entrainment velocity to the 0.7 power.
For typical industrial worm gear operating conditions, film thickness ranges from 0.3 to 1.5 µm.
The lambda ratio λ = h₀ / σ, where σ is the composite roughness of the two surfaces (σ = √(Ra₁² + Ra₂²)). For a worm at Ra 0.4 µm meshing with a wheel at Ra 0.8 µm, σ = √(0.16 + 0.64) = 0.89 µm. With film thickness 0.8 µm, lambda = 0.8 / 0.89 = 0.9, which is mixed lubrication regime.
Three regimes have very different consequences. Lambda greater than 3 (full EHL): the surfaces are completely separated, wear is governed by oxidation and additive depletion, service life is on the order of 50,000 to 100,000+ hours. Lambda 1 to 3 (mixed): partial metal contact, moderate wear, service life 10,000 to 50,000 hours. Lambda less than 1 (boundary): extensive metal contact, accelerated wear, service life 1,000 to 10,000 hours and risk of scuffing.
For most industrial worm gear specifications, the design target is lambda 1.5 to 2.5 — solidly in the mixed regime, with margin against boundary lubrication during cold starts and load excursions. Achieving this target typically means worm Ra 0.4 to 0.8 µm and wheel Ra 0.8 to 1.6 µm with appropriate viscosity oil. Specifying smoother finishes pushes lambda above 3 and into full EHL — useful for premium applications but not necessary for the broad middle of industrial demand.
A Korean dairy processor specified worm gear pairs for a yogurt cup-filling machine where the worm gear pair drove a metering screw in direct food contact. Regulatory requirement: Ra ≤ 0.4 µm on all food-contact surfaces per 3-A Sanitary Standards. Standard hobbed-and-ground worm at Ra 0.6 µm did not meet specification. Engineering specified hobbed plus ground plus polished worm at Ra 0.2 µm and AISI 316L stainless wheel at Ra 0.3 µm. Cost premium over standard ground: 320 USD per pair (about 2.0× the standard ground price). The 320 USD premium was non-negotiable; without it, the equipment could not be sold to the Korean dairy market at all. Field service over 3 years: zero surface-related failures, zero regulatory citations, full pass on annual sanitary audit. Lesson: regulated industries (food, pharma, sterile) make the surface finish decision regardless of cost — specify to the regulation, period.
A Japanese rotary indexer builder specified worm gear pairs for 8-station precision machining centres with positioning accuracy requirement plus or minus 4 arcseconds. Standard ground worm at Ra 0.6 µm tested at lambda 0.85 with the application’s high-viscosity gear oil — borderline boundary regime. Lapped pair at Ra 0.25 µm pushed lambda to 1.6, into stable mixed regime. Cost premium: 420 USD per pair over standard ground (about 1.3× ground). Test bench results: wear rate at the lapped finish ran 0.8 micrometres of bronze removal per 1,000 hours of operation, against 3.4 micrometres per 1,000 hours for the ground-only finish. Projected service life ratio at 4× longer for the lapped pair, justifying the cost premium across the equipment 12-year service horizon. Decision: lapped finish, accepting 4-week additional lead time. Lesson: in precision applications where dimensional drift over service life matters, lapped finish protects the geometry stability that the customer paid for.
A Vietnamese conveyor manufacturer building light-duty parts conveyors evaluated worm gear surface finish options. Standard ground at Ra 0.6 µm quoted at 220 USD per pair. Hobbed-only at Ra 1.8 µm quoted at 145 USD per pair. The conveyor application ran 10 hours per day at 35 percent of rated capacity — well below the boundary lubrication risk threshold even at the rougher Ra. Engineering specified hobbed-only pair plus a run-in protocol (50 hours at 30 percent load, then 50 hours at 60 percent load before commissioning at full operation). Pair surface roughness measured after 100 hours of run-in: worm thread Ra 1.5 µm (essentially unchanged), bronze wheel Ra 0.55 µm (reduced from 1.8 µm during run-in). Operating lambda at the run-in steady state: 1.4. Cost saving against ground specification: 75 USD per pair × 240 unit annual production = 18,000 USD per year. Field reliability over 3 years: average pair service life 7.2 years, exceeding the 6-year target. Lesson: for moderate-duty applications, hobbed-only with a defined run-in protocol delivers reliable service at significantly lower cost than ground specification. Browse wormwielreductor options where surface finish is specified at the right tier for the duty class — not over-specified by default.
All three parameters describe surface roughness but emphasise different features. Ra (arithmetic average) is the average absolute deviation from the mean line — most commonly specified. Rz (mean roughness depth) is the average peak-to-valley distance over five sampling lengths — sensitive to occasional defects. Rmax is the largest peak-to-valley distance in the sampling length — most sensitive to single defects. For worm gear specifications, Ra is the standard call-out. Rz is added when localised defects matter (sanitary applications, high-precision indexers). Rmax is rare except in critical bearing or seal contexts. Typical relationships: Rz is roughly 4 to 7 times Ra; Rmax is roughly 1.2 to 1.5 times Rz.
Yes, on the bronze wheel specifically and only up to a point. Phosphor bronze peaks deform plastically and abrasively wear during the first 100-300 hours against the harder steel worm. Typical improvement: Ra 1.8 µm as-manufactured to Ra 0.5-0.8 µm after run-in. The steel worm does not change measurably. The effect is most pronounced when initial conditions favour mild wear (good lubrication, moderate load, controlled temperature) and least pronounced under aggressive conditions (boundary lubrication, shock load) where micro-pitting takes over before the run-in smoothing completes. Specifying a defined run-in protocol (typically 50 to 100 hours at 30-50 percent load) maximises the smoothing benefit and minimises the risk of premature wear.
Three potential downsides for superfinished worm gear surfaces. First, cost — typical 2.5 to 3 times standard ground pricing, which only justifies in regulated or premium applications. Second, the smoother surface offers less natural lubricant retention; the long-discredited “oil pocket theory” had merit at the extreme — Ra below 0.05 µm can show film starvation in some operating regimes. Modern superfinish specifications avoid this by targeting Ra 0.1 to 0.2 µm rather than going to the absolute minimum. Third, in non-pristine environments, debris and contamination preferentially abrade the smooth surface — a worm gear pair operating in a dusty foundry or cement plant gets faster wear from a superfinished worm than a ground worm because the smooth surface has no asperities to “absorb” small particles. For industrial applications where cleanliness control is realistic, superfinish is genuinely beneficial; for applications where it is not, ground finish gives more practical durability.
Three methods. Stylus profilometry is the standard: a diamond-tipped stylus traces across the surface, recording vertical deflection as a profile, from which Ra and other parameters compute. Used on dedicated profilometers (Mitutoyo Surftest, Mahr Perthometer, Taylor Hobson Talysurf) — measurement takes 30 seconds per trace, repeatability roughly plus or minus 5 percent. Optical profilometry uses focus-variation or interferometric techniques to scan the surface without contact — slower and more expensive but produces 3D surface maps useful for research. Atomic force microscopy reaches sub-nanometre resolution but is impractical for production inspection. For routine worm gear flank measurement, stylus profilometry is the universal standard, ISO 4287 specifies the procedure, and reputable suppliers include Ra measurement reports in standard documentation packages.
The active flank — the side that engages under operating load — sees full contact stress and full sliding velocity. This is where surface finish matters and where the Ra specification applies. The opposite flank engages only briefly during reverse rotation or backlash takeup, at low load. Specifying premium finish on both flanks adds cost without proportional benefit. Modern worm gear specifications distinguish between active-flank Ra (typically 0.4 to 0.8 µm for ground) and reverse-flank Ra (typically Ra 1.6 µm or as-hobbed). The cost saving from finishing only the active flank can be 20 to 40 percent of total finishing cost. For applications where reverse loading is significant (bidirectional drives, hoists, indexers with both directions), both flanks should receive the same finish.
Extreme-pressure (EP) additives in worm gear oils form chemical reaction layers on the metal surface during boundary contact. These layers protect against scuffing in the periods when the lubricant film is too thin to fully separate the surfaces. EP additives are most active at higher temperatures and require some boundary contact to activate. A worm gear pair operating in full EHL regime (lambda greater than 3) sees little EP additive activity because boundary contact rarely occurs. A pair in mixed regime (lambda 1-3) sees moderate EP layer formation. A pair in boundary regime needs maximum EP additive concentration. Surface finish therefore interacts with additive selection: smoother surfaces operate in cleaner EHL regime and need less aggressive EP package; rougher surfaces operate in mixed regime and need higher EP additive levels. Mismatching surface finish and oil grade is a common diagnostic finding for unexpectedly short worm gear service life.
No — they are different processes with different effects. Electropolishing is an electrochemical process that removes surface metal preferentially at high points, producing a clean smooth surface typically Ra 0.1 to 0.4 µm depending on substrate condition. It is most often used on stainless steel for sanitary applications. Mechanical polishing on a worm thread uses abrasives or vibratory media to physically remove peaks, producing similar Ra range but with slightly different surface morphology — directional polishing marks rather than the smoother random topography of electropolish. For worm gear food-contact applications, both processes meet typical Ra targets; for premium NVH or efficiency applications, mechanical polishing is more common because it preserves the precise tooth geometry better than the slightly material-removing electropolish process.
Worm gear surface finish is the friction language of every meshing pair — Ra and the resulting lambda ratio determine whether the lubricant film fully separates the surfaces (full EHL, long service life) or allows intermittent contact (mixed or boundary lubrication, accelerated wear). Four finish processes cover the practical range from hobbed-only at Ra 1.6 to 3.2 µm through to polished at Ra 0.1 to 0.2 µm, each step roughly halving roughness and doubling cost. The right finish for a given application follows from duty class, lubrication regime, and regulatory requirement — not from a default “smoother is better” preference. Most industrial worm gear pairs operate well at Ra 0.4 to 0.8 µm; precision indexers and high-power applications justify lapped or polished finishes; food and pharmaceutical applications mandate Ra ≤ 0.4 µm regardless of mechanical need. The practical insight is that the worm finish is permanent (hardened steel does not wear) while the wheel finish improves with run-in (soft bronze self-polishes during the first 100 to 300 hours). Specify the worm at the operating-regime target and accept the wheel finish one tier rougher; over-specifying the wheel wastes money the wheel will achieve naturally.
Send the application duty class, lubrication regime, and any regulatory requirements. We will recommend the right finish tier (hobbed, ground, lapped, or polished) with cost and lead time for each option — typically within one Korean working day for standard catalogue specifications.
Redacteur: Cxm
Worm and Worm Wheel Pair Matching — Why Mix and Match Fails A worm and…
Worm Gear Strength Calculation — DIN 3996, ISO 14521, AGMA 6034 From application torque to…
Worm Gear Contact Pattern — How Bluing Tests Reveal Quality A 60 to 80 percent…
Worm Gear Module — Choosing the Right Tooth Size for Torque What module do I…
Worm Gear Center Distance — How to Calculate and Standardise One millimetre of centre distance…
Worm Gear Tooth Profile — ZA, ZN, ZI, ZK and How to Choose Why is…