Gear Design and Power Transmission: The Complete Engineering Guide to Spur, Helical, Bevel, and Planetary Gear Systems

Complete engineering guide to gear design: involute geometry, Lewis bending and Hertzian contact stress equations, material selection from case-hardened steel to 3D-printed polymers, and practical design for spur, helical, bevel, and planetary gear trains.

·

Gear Design and Power Transmission: The Complete Engineering Guide to Spur, Helical, Bevel, and Planetary Gear Systems

Gears are among the oldest and most fundamental machine elements — and among the most misunderstood. A well-designed gear pair transmits power at 98–99.5% efficiency for thousands of hours. A poorly designed one strips its teeth in seconds. The difference comes down to understanding the geometry, the materials, and the forces involved.

This guide covers gear design from first principles: the involute curve that defines every tooth profile, the standardised metric and imperial systems, the bending and contact stress equations that determine whether a gear survives, and the practical differences between spur, helical, bevel, and planetary configurations. Throughout, we apply the analysis to both machined metal gears and 3D-printed polymer gears — because the same physics governs both, but the failure modes differ dramatically.

1. Why Gears: Torque, Speed, and the Fundamental Law

Every gear train exists for one of three reasons: to reduce speed and multiply torque (reducer), to increase speed (multiplier), or to change the axis of rotation. The fundamental law of gearing states that the angular velocity ratio between two meshing gears must remain constant throughout the mesh cycle — any deviation produces vibration, noise, and accelerated wear.

For two gears with tooth counts N_1 and N_2:

Where \omega is angular velocity (rad/s) and T is torque (N·m). The gear ratio i = N_2 / N_1. A 20:60 tooth pair gives i = 3, meaning the output shaft rotates at one-third the input speed but delivers (ideally) three times the torque, minus efficiency losses:

Typical efficiencies: spur gears \eta \approx 0.98–0.995 per stage, helical gears 0.98–0.99, worm gears 0.40–0.90 (strongly dependent on lead angle), planetary gear trains 0.97–0.99 per stage.

2. The Involute Tooth Profile: Why That Shape Works

2.1 The Geometry Problem

To maintain a constant velocity ratio, the tooth profiles must satisfy the condition that the common normal at every point of contact passes through a fixed point on the line of centres — the pitch point. The involute curve is the unique solution that also tolerates centre-distance variations without affecting the velocity ratio, making it the universal standard for power-transmission gearing.

2.2 Generating the Involute

An involute is the curve traced by a point on a taut string as it unwinds from a base circle. In parametric form, for base circle radius r_b:

Where \theta is the roll angle in radians. The pressure angle \phi at any point on the involute satisfies:

At the pitch circle (radius r), the pressure angle is standardised: 20^\circ in metric systems (ISO 53), 14.5^\circ or 20^\circ in imperial (AGMA). The 20^\circ standard replaced the older 14.5^\circ because it permits fewer teeth without undercutting — 18 vs 32 teeth minimum for full-depth teeth.

2.3 Key Circle Diameters

For a gear with module m (mm) and N teeth:

Parameter · Symbol · Formula

Pitch diameter · d · m \cdot N

Base circle diameter · d_b · d \cdot \cos\phi

Addendum circle · d_a · d + 2m

Dedendum circle · d_f · d - 2.5m

Circular pitch · p · \pi \cdot m

Tooth thickness (at pitch circle) · s · p/2 = \pi m/2

Centre distance · a · m(N_1 + N_2)/2

For imperial (diametral pitch P in teeth/inch): d = N/P, p = \pi/P, a = (N_1+N_2)/(2P).

The addendum and dedendum values (1.0m and 1.25m respectively) define the full-depth involute system per ISO 53. Stub teeth use 0.8m addendum and 1.0m dedendum for higher bending strength at the cost of contact ratio.

3. Contact Ratio: How Many Teeth Carry the Load

The contact ratio \varepsilon is the average number of tooth pairs in contact during the mesh cycle. It's the ratio of the path of contact length to the base pitch:

Where p_b = \pi m \cos\phi is the base pitch. For smooth operation, \varepsilon \geq 1.2 is recommended; below 1.0, the next tooth pair engages before the previous one disengages, producing shock loads and noise.

Practical rule: A 20-tooth, 20^\circ pressure angle spur gear pair at 1:1 ratio gives \varepsilon \approx 1.56. Increasing the pressure angle to 25^\circ reduces \varepsilon to ~1.35 (fewer teeth in contact) but strengthens the tooth root — a deliberate trade-off in aerospace and automotive gearing.

For helical gears, the total contact ratio adds an axial component:

Where \varepsilon_{axial} = \frac{b \sin\beta}{\pi m_n} (b = face width, \beta = helix angle, m_n = normal module). This is why helical gears run quieter — the axial overlap means at least 2–3 teeth are always engaged.

4. Forces on Gear Teeth

4.1 Spur Gear Forces

The normal force F_n acts along the line of action (at pressure angle \phi to the tangent). Resolving into tangential and radial components:

For a 20^\circ pressure angle, \tan 20^\circ = 0.364, so the radial (separating) force is about 36% of the tangential force. This radial force must be carried by the shaft bearings.

4.2 Helical Gear Forces

Helical gears add an axial thrust component due to the helix angle \beta:

Where \phi_n is the normal pressure angle (typically 20^\circ). The axial thrust F_a must be absorbed by thrust bearings or countered by using opposite-hand helical gears on the same shaft (herringbone or double-helical arrangement).

For \beta = 15^\circ, F_a \approx 0.27F_t — manageable with deep-groove ball bearings. For \beta = 30^\circ, F_a \approx 0.58F_t — requires tapered roller or angular-contact bearings.

5. Gear Strength: Lewis Bending and AGMA Equations

5.1 The Lewis Bending Equation (1892)

Wilfred Lewis modelled a gear tooth as a cantilever beam of parabolic profile, loaded at the tip. The bending stress at the root:

Where:

The Lewis form factor for 20^\circ full-depth teeth:

Teeth · 12 · 14 · 17 · 20 · 25 · 30 · 40 · 60 · 100 · Rack

Y · 0.245 · 0.276 · 0.303 · 0.322 · 0.345 · 0.359 · 0.383 · 0.414 · 0.446 · 0.484

The Lewis equation is the starting point — it ignores stress concentrations at the root fillet and dynamic effects. But it gives a quick go/no-go check: compare \sigma_b against the allowable bending stress of the material (with a safety factor).

Example calculation: A steel spur gear (m = 3 mm, N = 20 teeth, b = 24 mm) transmitting T = 15 N·m.

Against EN8 steel (yield ~350 MPa, endurance ~180 MPa): comfortable margin. For 3D-printed PLA (flexural strength ~60 MPa, but with significant notch sensitivity and layer adhesion weakness): the margin narrows considerably.

5.2 The AGMA Bending Stress Equation

The modern AGMA 2001-D04 standard refines Lewis with multiple correction factors:

Where:

The dynamic factor K_v deserves attention. For commercial-quality gears (AGMA Q8–Q9):

Where v is pitch-line velocity (m/s). For v < 10 m/s and Q9 quality, K_v \approx 1.1–1.3. At v = 0.5 m/s (typical for a 3D printer extruder gear at 100 RPM, 15 mm pitch diameter), K_v \approx 1.02 — dynamic effects are negligible. At v = 20 m/s (automotive transmission), K_v \approx 1.4–1.8 — significant.

5.3 Hertzian Contact Stress (Pitting Resistance)

Bending stress governs tooth breakage; contact stress governs surface pitting and spalling. The AGMA contact stress equation:

Where:

For steel-on-steel (E = 207 GPa, \nu = 0.3): Z_E \approx 191 \sqrt{\text{MPa}}. For steel-on-polymer (PLA: E \approx 3.5 GPa, \nu \approx 0.35): Z_E \approx 58 \sqrt{\text{MPa}} — much lower contact stress due to the compliant polymer conforming to the steel surface.

The allowable contact stress number S_c for common gear materials:

Material · S_c (MPa) · Hardness

EN8 / C45 (normalised) · 450–520 · 170–210 HB

EN24 / 4340 (hardened) · 860–1030 · 300–360 HB

20MnCr5 (case-hardened) · 1200–1350 · 58–62 HRC

42CrMo4 (induction-hardened) · 1000–1150 · 50–55 HRC

Cast iron (Grade 250) · 350–450 · 180–240 HB

POM (Delrin/acetal) · 25–40 · —

Nylon 6/6 (conditioned) · 15–25 · —

PLA (3D printed) · 4–12 · —

The 100:1 ratio between case-hardened steel and 3D-printed PLA contact stress capacity explains why polymer gears succeed only in low-load, low-speed applications — or as sacrificial elements designed to fail before expensive metal components.

6. Gear Types and Their Engineering Trade-offs

6.1 Spur Gears

Characteristics: Teeth parallel to axis of rotation. Simplest to manufacture (hobbing, shaping, 3D printing). Zero axial thrust. Contact is instantaneous along a line — each tooth pair engages and disengages abruptly, producing the characteristic gear whine.

Noise: The engagement frequency f = \frac{N \cdot \omega}{60} (Hz). A 20-tooth gear at 3000 RPM produces f = 1000 Hz — squarely in the most sensitive range of human hearing (1–4 kHz). This is why spur gears in automotive transmissions are being replaced by helical gears: not for strength, but for NVH (Noise, Vibration, Harshness).

Minimum teeth without undercutting: For 20^\circ full-depth: N_{min} = \frac{2}{\sin^2\phi} = \frac{2}{\sin^2 20^\circ} \approx 17.1 \approx 18 teeth. Below this, the involute profile is undercut by the generating cutter, weakening the tooth.

6.2 Helical Gears

Characteristics: Teeth cut at an angle \beta (helix angle, typically 15^\circ–30^\circ) to the axis. Engagement is gradual — contact starts at one end of the tooth and progresses diagonally across the face. This produces significantly less noise and vibration than spur gears.

The virtual number of teeth (for bending strength calculations in the normal plane):

For N = 20, \beta = 20^\circ: N_v \approx 24.1 — the tooth is effectively stronger because it's "longer" in the normal plane.

Efficiency penalty: The sliding velocity between helical teeth has an additional component along the tooth flank, increasing friction slightly. At \beta = 15^\circ, the efficiency loss is ~0.5% per stage. At \beta = 30^\circ, ~1–2% loss. Worm gears are the extreme case of helical gearing with \beta chosen for self-locking rather than efficiency.

Double helical / herringbone: Two opposite-hand helical gear halves on the same blank cancel axial thrust internally. Citroën famously used herringbone gears in the Traction Avant (1934); the double-helix became their logo.

6.3 Bevel Gears

Characteristics: Conical gears for intersecting shafts (typically 90^\circ). The tooth profile tapers from the outer (heel) to inner (toe) end. Manufactured on specialised Gleason or Klingelnberg bevel gear generators.

The pitch cone angle \gamma for the pinion (smaller gear):

Where \Sigma is the shaft angle (typically 90^\circ). For a 1:1 ratio at 90^\circ: \gamma_1 = \gamma_2 = 45^\circ.

Spiral bevel gears add a curved tooth path (spiral angle \psi \approx 35^\circ) for smoother engagement — the bevel equivalent of helical gears. They're ubiquitous in automotive differentials and helicopter main rotor transmissions. The Gleason system dominates in the US; Klingelnberg (palloid) in Europe.

Practical limitation: 3D printing bevel gears is challenging — the overhanging tooth profile on the conical surface requires support material that affects tooth surface quality. CNC machining or injection moulding is preferred for production.

6.4 Planetary (Epicyclic) Gear Trains

Characteristics: One or more planet gears revolve around a central sun gear, all contained within an internal ring gear (annulus). Three coaxial elements (sun, carrier, ring) provide two degrees of freedom — lock any one element to define the input/output relationship.

The fundamental planetary equation (Willis equation):

Where \omega_s, \omega_c, \omega_r are the angular velocities of sun, carrier, and ring; N_s, N_r are sun and ring tooth counts. The negative sign reflects the direction reversal through the planet gears.

Common configurations:

Configuration · Fixed · Input · Output · Ratio

Speed reducer · Ring · Sun · Carrier · 1 + N_r/N_s

Speed increaser · Ring · Carrier · Sun · \frac{1}{1 + N_r/N_s}

Reverse · Carrier · Sun · Ring · -N_r/N_s

Direct drive · — · Any two locked together · Third · 1:1

For N_s = 20, N_r = 80 (3 planets): the reducer ratio is 1 + 80/20 = 5:1. Stacking two stages gives 25:1 in the same axial envelope as a single-stage spur reducer giving 5:1 — the fundamental appeal of planetary gearing for compact high-ratio drives.

Load sharing: With n equally spaced planets, the tangential load divides as F_t/n — but only if the carrier is perfectly rigid and the planets are precisely positioned. In practice, a floating sun gear or flexible planet pins compensate for manufacturing tolerances. Without load sharing, one planet carries 60–70% of the load and fails prematurely.

7. Backlash: The Necessary Evil

Backlash is the clearance between mating tooth flanks, measured as the lost angular motion when reversing direction. It's necessary to accommodate:

Typical backlash ranges:

Anti-backlash designs: Split gears with spring-loaded halves, dual-pinion preloaded drives, and harmonic drives (strain wave gearing) achieve effectively zero backlash. The trade-off is increased friction, reduced efficiency, and higher cost.

For 3D-printed gears, a practical starting point is 0.3m circumferential backlash — so a module 2 mm gear pair gets ~0.6 mm backlash. Iterate from there based on trial fitting.

8. Gear Materials: Steel, Brass, Polymer, and 3D-Printed

8.1 Steel Gears

The workhorse material. Selection hierarchy:

Through-hardened steels (EN8/C45, EN24/4340): 170–360 HB. Used for medium-duty industrial gears. Through-hardening limits the gear size: sections >50 mm may not harden uniformly. Surface durability is moderate — pitting resistance limited to ~500–1000 MPa contact stress.

Case-hardened steels (20MnCr5, 8620, 17CrNiMo6): Carburising produces a high-carbon case (0.7–1.0% C, 58–62 HRC) with a tough, low-carbon core. Case depth typically 0.1–0.2 × module, or 0.5–2.0 mm for modules 3–10. The compressive residual stress in the case improves bending fatigue by 30–50% over through-hardened gears.

Nitrided steels (31CrMoV9, Nitralloy 135M): Lower distortion than carburising — suitable for precision gears that can't be ground after heat treatment. Surface hardness 900–1100 HV (~67 HRC equivalent), but the case is thinner (0.3–0.5 mm) and the hardness drops off steeply.

8.2 Brass and Bronze

Brass (C36000): Excellent machinability. Used for low-load instrument gears, clock mechanisms, and model engineering. Corrosion-resistant but soft: S_c \approx 80–120 MPa.

Phosphor bronze (C51000): The standard worm wheel material. The bronze sacrifices itself to protect the steel worm — the conformable bronze surface embeds debris and wears preferentially. Also used in high-load plain bearings.

8.3 Engineering Polymers

POM (Delrin/acetal): The default polymer gear material. Self-lubricating, excellent fatigue resistance, absorbs ~0.2% moisture (dimensional stability). Tensile strength ~70 MPa, flexural modulus ~2.8 GPa. Widely used in office equipment, automotive actuators, and consumer products. Can be injection-moulded to AGMA Q10 accuracy without post-processing.

Nylon 6/6 (PA66): Tougher than POM but absorbs 2–3% moisture (dimensions change with humidity). Often used with glass-fibre reinforcement (PA66-GF30) for strength and dimensional stability. Common in power tools and automotive window regulators.

PEEK: The high-performance option. Continuous service at 250°C, tensile strength ~100 MPa, flexural modulus ~4 GPa. Costs ~₹8,000–₹12,000/kg vs ~₹400/kg for POM. Used in aerospace, medical devices, and semiconductor equipment where failure is not an option.

8.4 3D-Printed Gears

Material · Flexural Strength (MPa) · Wear Resistance · Printability · Notes

PLA · 60–70 · Poor (brittle) · Excellent · Use for prototypes, low-load demonstrations. Not for sustained use — creep under load.

PETG · 50–65 · Moderate · Good · Better toughness than PLA. Moderate creep resistance. Decent for low-speed functional gears.

ABS · 45–60 · Moderate · Moderate (warping) · Better temperature resistance (85°C vs 55°C for PLA). Acetone smoothing improves surface finish.

Nylon (PA12) · 50–70 · Good · Difficult (warping) · Excellent wear resistance. SLS-printed PA12 is the best 3D-printed gear material.

Resin (Standard) · 40–60 · Poor (brittle) · Excellent detail · Very smooth surface, low friction. Brittle — fails catastrophically.

Resin (Tough/ABS-like) · 50–70 · Moderate · Good · Better impact resistance. Siraya Tech Blu or Formlabs Tough 2000 recommended.

PEEK (industrial FDM) · 100–120 · Excellent · Requires 400°C+ hotend · Best performance but limited to high-end printers (Intamsys, Roboze).

Design rules for 3D-printed gears:

  1. Orient teeth in the XY plane. Layer lines parallel to tooth contact surfaces are weakest — the tooth can delaminate under load. Print gears flat (axis vertical) so loads are perpendicular to layer lines.
  1. Minimum 4 perimeters. Tooth flanks are thin cross-sections. Four perimeters ensure the tooth is solid plastic, not sparse infill. At 0.4 mm nozzle, 4 perimeters = 1.6 mm solid wall.
  1. Root fillet radius ≥ 0.5× module. The stress concentration at the tooth root is the dominant failure site. A generous fillet (≥ 1 mm for m = 2) dramatically improves bending fatigue life. Most CAD gear generators default to near-zero fillet — overriding this is essential.
  1. Lubrication helps enormously. A thin application of PTFE or silicone grease on PLA/PETG gears reduces friction coefficient from ~0.4 (dry) to ~0.1–0.15, decreasing contact stress and wear rate by 3–5×.

9. Gear Manufacturing Methods

9.1 Hobbing

A hob (helical cutting tool with rack-form teeth) rotates in mesh with the gear blank, generating the involute profile kinematically. One hob can cut any tooth count of the same module and pressure angle — the generating process automatically produces the correct involute.

Accuracy: AGMA Q8–Q10 achievable. Production rate: 30–120 seconds per gear (module 2–5, 20–60 teeth). The most common production method for external spur and helical gears.

9.2 Shaping

A pinion-shaped cutter reciprocates axially while rotating in mesh with the blank. The only practical method for cutting internal gears (ring gears). Also used for shoulder gears where the hob can't run out.

9.3 Gear Grinding

After heat treatment, gear teeth distort (typically 20–50 µm). Grinding with formed or generating wheels restores accuracy to AGMA Q12–Q14 (1–3 µm profile error). Essential for aerospace and high-speed gears (>30 m/s pitch-line velocity).

9.4 Wire EDM

For one-off and prototype gears, wire EDM can cut hardened steel blanks to AGMA Q10 accuracy with zero tooling cost. Maximum thickness ~200 mm. Surface finish ~0.8–1.6 µm Ra. The heat-affected zone is negligible (5–20 µm). An excellent option for custom gears when volume doesn't justify a hob.

9.5 Injection Moulding

Polymer gears are moulded to AGMA Q9–Q10 accuracy straight from the mould — no secondary finishing. The mould cavity is spark-eroded into hardened tool steel ($60,000–₹200,000 for a single-cavity gear mould). Cycle time 10–30 seconds. At 100,000+ pieces, injection moulding is the only economically viable option.

9.6 3D Printing

FDM gears are AGMA Q6–Q8 at best (tooth-to-tooth error 100–200 µm). SLA/DLP can reach Q8–Q9 (50–100 µm). SLS PA12 can reach Q9–Q10 (25–50 µm). The surface roughness of FDM (10–25 µm Ra layer lines) acts as a wear accelerant — mating printed gears essentially lap each other into shape, producing plastic dust.

10. Lubrication and Wear Mechanisms

10.1 Elastohydrodynamic Lubrication (EHL)

At the tooth contact, the lubricant experiences pressures of 500–2000 MPa. At these pressures, the oil viscosity increases by 3–6 orders of magnitude and the steel surfaces deform elastically. The result is an EHL film, typically 0.1–1 µm thick — thinner than the surface roughness of all but the finest-ground gears. Most gear contacts operate in the mixed lubrication regime: partial EHL film, partial asperity contact.

The specific film thickness \lambda:

Where h_{min} is the minimum EHL film thickness and R_q is the RMS surface roughness. For \lambda > 3: full EHL, negligible wear. For 1 < \lambda < 3: mixed, moderate wear. For \lambda < 1: boundary lubrication, high wear, risk of scuffing.

10.2 Failure Modes

Mode · Mechanism · Appearance · Prevention

Bending fatigue · Cyclic root stress → crack → fracture · Clean fracture at root, beach marks · Increase module, shot peening

Pitting · Contact fatigue → subsurface cracks → surface spalls · Small craters on pitch line · Higher-viscosity oil, case hardening

Scuffing · Oil film breakdown → metal-to-metal welding and tearing · Scored, torn surface in sliding direction · EP additives, lower contact stress

Abrasive wear · Hard particles in lubricant or from tooth surface · Polished, matte surface · Filtration, hardened surfaces

Plastic flow · Overload → yielding of surface layer · Ridges at pitch line, "cold flow" · Reduce load, harder material

10.3 Lubricant Selection

11. Designing a Gear Pair: A Worked Example

Let's design a speed-reducing gear pair for a small CNC spindle drive: input 3000 RPM, 0.5 N·m torque; output target 1000 RPM, with 20 mm centre distance constraint.

Step 1 — Ratio: i = 3000/1000 = 3. Choose N_{pinion} = 20, then N_{gear} = 60. (Centre distance will determine module.)

Step 2 — Module: a = m(N_1+N_2)/2 = m(20+60)/2 = 40m. With a = 20 mm: m = 0.5 mm. This is fine for a low-torque application.

Step 3 — Checks:

Step 4 — Bending stress (pinion):

For 3D-printed PETG (flexural strength ~60 MPa, with safety factor 2 → allowable 30 MPa): FAILS. The gear needs a larger module or wider face.

Step 5 — Redesign: Increase module to m = 1 mm. Centre distance becomes 40 mm — exceeds constraint. Compromise: use N_{pinion} = 15 (below 18 risks undercutting — use profile shift), N_{gear} = 45, m = 1 mm → a = 30 mm. Face width b = 10 mm.

Now within PETG's allowable stress with a comfortable safety margin.

Step 6 — Profile shift: To avoid undercutting at N = 15, apply a positive profile shift x = \frac{17 - N}{17} = 0.118 (for 20^\circ pressure angle). This moves the cutter outward, thickening the tooth root. Centre distance increases slightly: a_{actual} = a_{nominal} + 2xm = 30 + 2(0.118)(1) = 30.24 mm.

This example illustrates the iterative tension in gear design: module, tooth count, centre distance, material strength, and manufacturing constraints are all coupled. Software tools (KISSsoft, MITCalc, or even the involute gear generator in Fusion 360) make the iteration manageable.

12. Gear Standards Summary

Standard · Region · System · Pressure Angle · Notes

ISO 53 / DIN 867 · International · Module (mm) · 20^\circ · Full-depth (h_a = m, h_f = 1.25m)

AGMA 201.02 · USA · Diametral pitch (\text{in}^{-1}) · 20^\circ / 25^\circ · Coarse pitch: P \leq 20; Fine: P \geq 20

JIS B 1701 · Japan · Module (mm) · 20^\circ · Similar to ISO, slight differences in addendum modification

BS 436 · UK (legacy) · Module or DP · 20^\circ · Superseded by ISO

The key takeaway: 20^\circ pressure angle, full-depth involute to ISO 53 is the de facto global standard. Unless you have a specific reason to deviate (aerospace 25^\circ, instrument 14.5^\circ for finer pitch), use 20^\circ full-depth.

13. Software Tools for Gear Design

Free/Open Source:

Commercial:

Online generators:

14. The Future: 3D-Printed Metal Gears and Topology Optimisation

Additive manufacturing is changing gear design in two ways:

L-PBF (Laser Powder Bed Fusion) steel gears can now achieve >99.5% density with mechanical properties approaching wrought material. The benefit isn't just prototyping — AM enables internal cooling channels, lattice structures for weight reduction, and custom tooth profiles that can't be hobbed. GE Aviation's Advanced Turboprop engine uses an AM gearbox with 30% fewer parts than conventional designs.

Topology optimisation removes material from non-load-bearing regions of gear blanks. A study by the Gear Research Centre (FZG, TU Munich) demonstrated 40% mass reduction in a planetary carrier with equivalent stiffness — reducing inertia and improving the transmission's response time.

For the maker and small-batch manufacturer, the immediate opportunity is hybrid gear trains: 3D-printed polymer planet gears meshing with steel sun and ring gears. The polymer gears act as mechanical fuses — they're intentionally the weakest link, protecting the more expensive metal components from shock loads. This pattern is already used in some cordless power tool transmissions.

Conclusion

Gear design is one of those disciplines where the physics hasn't changed in a century — the Lewis equation dates from 1892, and involute geometry from Euler in the 1760s — but the tools, materials, and manufacturing methods have transformed what's practical for the individual engineer.

If you take away three things from this guide:

  1. 20^\circ module-based involute is the default. Don't reinvent the wheel unless you have a specific reason.
  1. The Lewis equation is your first sanity check. If bending stress exceeds the material's allowable, no amount of lubrication or profile optimisation will save the gear — increase the module or widen the face.
  1. 3D-printed gears work — with caveats. Print flat (axis vertical), use 4+ perimeters, add generous root fillets, and lubricate. PLA is for prototypes; PETG or SLS nylon for functional parts.

Every robot arm, every CNC spindle, every 3D printer extruder is a gear design problem solved by someone who understood these principles. Now you do too.

Further Reading:

More FabFlow blog posts