Gear design calculator: spur, helical, bevel and worm gears, and a check of your Ansys gear results
Give the module, teeth, pressure angle, helix angle, profile shift, face width and the torque. You get the full geometry, the working pressure angle and centre distance, the contact ratio, undercut and interference checks, a drawing of the meshing pair from the true involute, the tooth forces, Lewis bending and Hertz contact stress with safety factors. Bevel and worm gears have their own tabs. Free, runs in your browser.
The fourth tab checks the spur, bevel and worm gear challenges of Skill-Lync's Ansys Workbench course: why the spur gear "deforms" 30 mm in every material, what stress to expect, which material to pick, and whether a 4, 5, 6 mm mesh study has converged. Skill-Lync read 697 public gear reports to build it.
KHK Gear Technical Reference · Shigley's Mechanical Engineering Design · Bhandari · Celik et al. 2008 · checked against a separate solution (1,526 checks at build) · Updated 22 Nov 2026
The 13 + 13 module-2 pair is what the Week 1 reports describe: 13 teeth is stated, module 2 is inferred from the 30 mm, and the 20 mm face width is assumed. Type your model's numbers. Helical: the module is the normal module and x is the normal shift.
Gleason straight bevel proportions (KHK Table 4.16). Forces at the mean diameter; bending and wear strength the Bhandari way, with the tangential force at the large end.
Axial module system, as KHK's Table 4.23. The friction curve is AGMA's, for a case-hardened steel worm on a bronze wheel; it is in ft/min in Shigley, converted here.
The materials are the values many Week 1 reports typed. Every result below changes with your inputs.
- Undercut: the pinion (13 teeth) needs x ≥ 0.24.
| Geometry | Pinion | Gear |
|---|---|---|
| Reference diameter d | 26 mm | 26 mm |
| Tip diameter da | 30 mm | 30 mm |
| Root diameter df | 21 mm | 21 mm |
| Base diameter db | 24.43 mm | 24.43 mm |
| Shift needed to avoid undercut, xmin | 0.24 | 0.24 |
| The pair | |
|---|---|
| Contact ratio εα | 1.442 |
| Forces and strength | Pinion | Gear |
|---|---|---|
| Lewis bending stress | 73.68 MPa | 73.68 MPa |
| Hertz contact stress | 816.4 MPa | same |
| Geometry | Pinion | Gear |
|---|
| Forces and strength | Pinion | Gear |
|---|
| Geometry |
|---|
| Running |
|---|
| Starts | Ratio | Lead angle | η, worm drives | η, wheel drives |
|---|
Same q, module and friction for 1, 2 and 4 starts.
| Material (E, ν) | Lewis root | Yield / Lewis | Hertz (× stiffest) | Yield / Ansys peak |
|---|
4. Has the mesh converged? (Celik et al. grid convergence index)
| Celik et al. procedure |
|---|
With more than three meshes the three finest are used. Element size, or N−1/3 from element counts in 3-D.
How do you calculate spur and helical gear dimensions?
From the module, the teeth and the basic rack. On the standard basic rack, as KHK tabulates it, the addendum is 1.00 m, the dedendum 1.25 m, the tooth depth 2.25 m and the root fillet 0.38 m. A helical gear uses the same rack in the normal plane, so its transverse module and pressure angle grow with the helix angle:
Profile shift moves the cutting rack out by x m. Shifted gears no longer mesh on their reference circles: they mesh at a working pressure angle αw found from the involute function, and the centre distance and addenda follow (KHK Tables 4.3 and 4.9):
The calculator solves the involute equation by Newton's method; the verifier solves it separately by bracketing. On KHK's shifted pair (module 3, 12 and 24 teeth, x 0.6 and 0.36) both give αw = 26.0886° and a = 56.4999 mm, as KHK prints. Pick the module from the standard series in the first box (JIS B 1701-2 as KHK lists it; series I first), or type any value.
What is the contact ratio, and how many teeth avoid undercut?
The contact ratio is the length of the path of contact divided by the base pitch: how many tooth pairs share the load on average. Keep it above about 1.2. The path runs along the line of action between the two tip circles, and a helical gear adds an overlap ratio across the face:
Skill-Lync's verifier checks this formula against a second, numerical route: it turns a parametric involute through the mesh, finds by root-finding the angle at which each point's normal passes through the pitch point (the law of gearing), and measures the turn over which contact stays inside both tip circles. On the 6 pairs tested, including KHK's, the two agree to within 10−5 wherever no tip passes the tangent point.
With few teeth, the rack cuts away the root of the involute. The limit and the shift that avoids it (KHK Eq. 3.7 and 3.10):
KHK accepts 14 teeth with slight undercut. The numerical check also catches interference that the formula hides. KHK's standard 12 + 24 pair has the gear's tip reaching past the tangent point T₁, below the pinion's base circle, where there is no involute: the formula gives 1.51, but contact on the involutes stops at T₁, so the real ratio is 1.41. With KHK's +0.6 / +0.36 shift the interference goes and the ratio is 1.2. Shigley's limits agree with this test: 13 teeth mesh with up to 16, and 12.3 teeth is the least for an equal pair at 20°.
The drawing is made from the same numbers: each flank is the true involute from the base circle to the tip,
Below the base circle the flank is drawn as a radial line, joined to the root circle by an arc of radius 0.38 m. The real root is a trochoid cut by the tool tip, which this does not model. The CSV buttons export one tooth as x, y points in mm, tooth centred on the y axis, for a sketch in SpaceClaim, CATIA or SOLIDWORKS (fit a spline through the flank points).
How do you check a spur gear for bending strength?
Compare the Lewis bending stress, or the load, with what the material allows. The tooth is a cantilever carrying the tangential force at its tip. Shigley writes the stress, and Bhandari's beam strength is the same equation solved for the load, with σb = Sut/3:
Y is from Shigley's Table 14-2 for 20° full-depth teeth, interpolated between rows, and read at the virtual tooth count z/cos³β for helical gears. The velocity factor raises the load for pitch-line speed V in m/s (Shigley Eq. 14-6 and AGMA Eq. 14-27):
Bhandari writes the first three as divisors, Cv = 3/(3 + v), 6/(6 + v) and 5.6/(5.6 + √v), and assigns them by how the teeth are made. The Shigley preset reproduces Example 14-1: a 16-tooth, 8-pitch AISI 1020 pinion at 1,200 rpm carrying 365 lbf gives 69.1 MPa, the 10 kpsi (68.9 MPa) the example allows.
How do you check gear teeth for wear and contact stress?
Two ways, and the calculator gives both. Bhandari's (after Buckingham) is a wear strength, the load at which the teeth start to pit:
K = 0.16 (BHN/100)² is for a steel pair at 20°; the calculator uses the softer gear's hardness. Shigley's is the Hertz contact stress of two cylinders with the involute radii at the pitch point (Eq. 14-13 and 14-14):
It is the same line-contact solution as the Hertz contact stress calculator, and the code is shared. For steel on steel ZE is about 190 √MPa (Shigley's Table 14-8 lists 191). For shifted gears the calculator uses the working pitch circle and αw; for helical gears it applies the spur formula in the transverse plane, which ignores the longer contact lines and overstates the stress.
How do you design a straight bevel gear?
Work on the pitch cones. The cone angles come from the teeth and the shaft angle, the teeth shrink towards the apex, and each tooth behaves like a spur tooth on the back cone (Tredgold's approximation):
The tooth proportions are KHK's Gleason table: working depth 2 m, tooth depth 2.188 m, the pinion's addendum longer than the gear's. On KHK's example (module 3, 20 and 40 teeth, b 22 mm) the calculator gives δ₁ 26.56505°, Re 67.08204 mm and the tip, root and crown dimensions KHK prints. Strength follows Bhandari: Lewis with the bevel factor 1 − b/Re, the tangential force at the large end, and wear strength 0.75 b Q d₁ K/cos δ₁ with Q from the virtual teeth. His Example 19.4 (20 and 30 teeth, module 4, 400 HB) checks out: 1,194 N at the large end and a wear strength of 5,112 N (he prints 5,113.53 from rounded inputs).
How efficient is a worm gear, and when is it self-locking?
Efficiency rises with the lead angle and falls with friction. The geometry follows KHK's axial-module table and the efficiency and forces follow Shigley (Eq. 13-43 to 13-46):
Friction falls as the sliding speed rises. Shigley gives AGMA's curve in ft/min, which the calculator converts:
KHK's 2-start worm (module 3, d₁ 44 mm, 30 teeth) at 1,440 rpm slides at 3.35 m/s, so μ ≈ 0.0254 and η = 83.2%. A single-start worm with q 12 needs μ ≥ 0.0783 to lock; at 960 rpm the curve gives 0.0296, so the wheel can drive it back at 62%. At 10 rpm a q 10 single start sees μ 0.11 and does lock. Do not use a worm as a brake: running friction is low and vibration lowers it further. The verifier re-derives the efficiency from a force and power balance on the thread and reproduces Shigley's Table 13-6 (f = 0.05) at αn 14.5°: every row except 5°, where Shigley's 9th edition prints 62.0% and the formula gives 62.6% (the SI 11th edition prints 62.6). (1 − η) of the input power heats the oil, which is why worm boxes are often rated by the heat their housing can shed.
Why is total deformation 30 mm in the Ansys spur gear challenge?
Because it is the rotation you applied, not elastic deformation. The Week 1 challenge turns one gear 180° in six steps. Ansys measures total deformation from each node's original position in the global coordinate system, so a point at radius r that turns through θ shows
At 180° that is the diameter. 6 reports state the model has 13 teeth per gear; 30 mm after 180° means a point at 15 mm, which is consistent with the tip radius of a 13-tooth gear of module 2 (da = m(z + 2) = 30 mm). That is an inference: no report text gives the module. The 30 mm is the same for every material because E does not enter a rigid turn. The bevel challenge does the same thing: 47.87 mm at all three meshes in a public report is a 120° turn of a point at 27.64 mm.
To see the deformation of the teeth instead, look at equivalent elastic strain, or use the Relative Deformation result of Ansys's Custom Results add-on, which removes the rigid-body motion; Ansys describes it as isolating "true material deformations from overall system movement" (2025 R2 help). The fourth tab flags any reported value within 2% of the rigid travel.
Which material is best for the spur gear in the Ansys challenge?
The one with the highest yield over its stress, not the one with the lowest stress. With a moment on one gear, statics fixes the tooth force: 769 N on the 26 mm pitch circle for 10 N m. Lewis bending stress does not depend on E, so it is the same in all three materials. Contact stress goes with √E*, so the softer material shows less of it:
| Material (values typed in many reports) | E, ν | Lewis root stress | Yield / Lewis | Hertz at the pitch point |
|---|---|---|---|---|
| Ductile cast iron | 171 GPa, 0.275 | 73.7 MPa | 4.67 | 736 MPa (0.911×) |
| Cast steel | 203 GPa, 0.3 | 73.7 MPa | 4.74 | 808 MPa (1×) |
| Cast bronze | 80 GPa, 0.345 | 73.7 MPa | 1.95 | 516 MPa (0.638×) |
13 + 13 teeth, module 2, 20°, face width 20 mm (assumed), 10 N m. At 10 mm the root stress doubles to 147 MPa and the contact stress rises by √2.
Cast bronze shows 36% less contact stress than cast steel and the same root stress, but its yield is 144 MPa against 349, so its margin is the smallest. Read the stress against each material's strength. The contact half-width is 32.2 to 50.5 µm, about thirty times smaller than a 1 mm element, so a 1 mm mesh cannot resolve the contact peak and will report less than the Hertz value.
What stress should the Ansys gear challenge give?
At the root fillet, close to the Lewis value: about 73.7 MPa for a 20 mm face, 147 MPa for 10 mm, with the 10 N m moment. At the contact, the Hertz value above. An FE peak elsewhere is usually a contact edge, the end of a tooth face or a sharp corner. Ansys's own guidance is that a stress singularity "will not converge to a specific value": refine the mesh and it keeps rising, while the stress nearby converges. Read the root fillet stress, compare it with Lewis, and treat a peak that grows with every refinement as a singularity, not a result. Type your peak into the fourth tab to see it against both hand values and each yield.
Has the bevel gear mesh converged at 4, 5 and 6 mm?
Usually not in the sense Celik et al. define. Their procedure needs three meshes refined by a ratio above 1.3; 6 to 5 to 4 mm is 1.2 and 1.25. It gives an observed order p, an extrapolated value and a grid convergence index:
The verifier reproduces Celik's Table 1: p 1.53 and GCI 2.2% for the reattachment length. Two public bevel reports show what the 4/5/6 mm study gives. One series (327.21, 399.94, 407.95 MPa at 6, 5, 4 mm) converges monotonically, but with p = 12.3, far above the 2 that quadratic elements can give, so its small GCI (0.17%) proves nothing. The other (5.4972, 4.4421, 5.5566 MPa at 4, 5, 6 mm) oscillates. Several public reports also put a 1.75 mm face sizing on the teeth in all three runs, so the mesh where the stress peaks does not change with the global size. Refine the contact region itself, by a ratio of at least 1.3.
Why does the worm gear show over 1,000 MPa with structural steel?
Because a linear elastic model has no yield. Structural steel's yield is 250 MPa; a linear model lets stress rise without limit, so a value above yield is not a real stress. The material would yield and spread the load. Look at where the peak sits: at a contact edge or a corner it is a local peak, and a frictionless contact may also let the surfaces penetrate. Check the penetration, refine the contact, add a plastic material model if the load is real, and read the stress away from the contact edge. The worm is turned 750°, so its total deformation is again mostly rigid travel: the decoder gives the radius that explains your number.
What do the gear reports on Skill-Lync show?
Skill-Lync read all 697 public reports on the three challenges on 22 Nov 2026 (275 spur, 247 bevel, 175 worm, after removing re-posts). Most results sit in screenshots, so each count is over the reports that state the number in their text, and is a floor.
- Spur, deformation: 58 reports give a total deformation in the text; 54 of them give 30 mm, and 21 give the same value for all three materials. Only 5 of 275 mention rigid motion. Example: report.
- Spur, material choice: of 100 conclusions that pick one, 58 pick cast steel, 25 ductile cast iron and 17 bronze. 68 compare the stress with a yield, ultimate strength or safety factor; 48 say the stress is about the same in every material; 8 pick the lowest stress. The E values typed range from 80 to 210 GPa.
- Spur, stress: 38 of the 40 that state a stress quote more than 100 MPa for the 10 N m moment.
- Bevel, mesh study: 168 use the 4/5/6 mm sizes and 114 call the result converged or mesh-independent; 1 uses Richardson extrapolation or a GCI, 10 mention a singularity and 3 rigid motion. Only 12 state three or more stress values in the text.
- Worm: 63 state a maximum equivalent stress; 60 of them are above the 250 MPa yield of structural steel (values such as 889, 1,178, 1,306, 1,620 and 3,055 MPa). 17 mention yield and 21 plasticity; 22 state the 750° rotation; total deformation is given as about 7.25 to 7.28 mm in 21 and 14.4 to 14.8 mm in 16. 5 mention self-locking and 14 efficiency.
What this calculator does not cover
- The full AGMA 2001 or ISO 6336 ratings. No geometry factor J, load distribution, size, rim or life factors: the bending check is Lewis with a velocity and application factor, and the contact check is Hertz at the pitch point.
- The true root fillet. The drawing and CSV use an arc of 0.38 m, not the trochoid the cutter generates, and the contact ratio assumes a full involute even where the gear is undercut.
- Internal gears, racks, spiral bevel and crossed helical gears, and the worm's bending and wear ratings, which Bhandari reads from charts.
- Allowable contact stress for ductile iron, cast steel and tin bronze: no open source Skill-Lync could read gives one, so those contact safety factors are left blank. The bending allowable is Sut/3 for every material.
Running a gear contact model is the easy part; knowing what stress and deformation it should give before you press solve is what CAE teams hire for. The spur, bevel and worm gear challenges are part of Skill-Lync's Ansys Workbench course.
Frequently asked questions
How do you calculate the centre distance of two gears?
For standard gears, a = (z1 + z2) m / 2: 54 mm for 12 and 24 teeth at module 3. With profile shift, solve inv αw = inv α + 2 tan α (x1 + x2)/(z1 + z2), then a = ((z1 + z2)/2 + y) m. KHK's example with x1 0.6 and x2 0.36 gives 26.09° and 56.4999 mm.
What is the minimum number of teeth on a spur gear to avoid undercut?
17 at a 20° pressure angle with no profile shift: 2/sin²α = 17.1, and KHK accepts 14 with slight undercut. A 13-tooth gear, like the Ansys Week 1 model, needs a profile shift of at least x = 1 − z sin²α/2 = 0.24 to cut cleanly. Undercut weakens the tooth root and shortens contact.
What is the Lewis equation for gear tooth bending?
σ = Kv Ft/(b m Y): tangential force Ft, face width b, module m and the Lewis form factor Y (Shigley Table 14-2: 0.261 for 13 teeth, 0.322 for 20, 0.485 for a rack). Kv raises the load for speed. Bhandari's beam strength, m b σb Y with σb = Sut/3, is the same equation as a load.
How do you calculate worm gear efficiency?
η = (cos αn − μ tan γ)/(cos αn + μ cot γ), with γ the lead angle, tan γ = z1/q. More starts mean a bigger lead angle and a higher efficiency. KHK's 2-start worm (γ 7.77°) at 1,440 rpm slides at 3.35 m/s; with Shigley's friction curve (μ 0.0254) it runs at 83.2%.
Is a single-start worm gear self-locking?
Only if the friction coefficient is at least cos αn tan γ. A single-start worm with q 12 has γ 4.76° and needs μ ≥ 0.0783. Running at 960 rpm, Shigley's curve gives μ 0.0296, so it back-drives at 62%. Crawling at 10 rpm, friction is higher and a q 10 single start locks.
Why is total deformation 30 mm in the Ansys spur gear challenge?
It is the rotation, not deformation. Ansys measures total deformation from the original position, so rigid motion counts. A point at radius r turned 180° moves 2r sin(θ/2) = 2r, so 30 mm means r = 15 mm, consistent with the tip of a 13-tooth module-2 gear. 54 of 58 reports that quote it give 30 mm.
Related on Skill-Lync
- Design of Machine Elements: spur gears, Lewis equation, dynamic load and wear
- Design of Machine Elements: helical, bevel and worm gears
- Theory of Machines: law of gearing, involute teeth and contact ratio
- Hertz contact stress calculator: point, line and elliptical contact
- Gear speed calculator: engine rpm and speed in each gear of a gearbox
- Planetary gear calculator: ratios, speeds and torques of an epicyclic set
- Mesh quality calculator
- Student project: Week 1 spur gear challenge
- Ansys courses compared
- All free engineering tools
Sources
- KHK Gear Technical Reference (KHK Gears, khkgears.net): involute gear profile (Table 3.1 basic rack, Table 3.2 module series from JIS B 1701-2:1999, Eq. 3.6-3.10 undercut, Table 3.4-3.5), calculation of gear dimensions (Tables 4.1, 4.3, 4.4, 4.9, 4.10 spur and helical, 4.16 Gleason straight bevel, 4.23 axial module worm pair, Eq. 4.27 self-locking), gear forces (Table 12.7). B. Dengel, "The importance of contact ratios", Gear Solutions (KHK-USA, March 2021): contact ratio tables.
- R. G. Budynas, J. K. Nisbett, Shigley's Mechanical Engineering Design, 9th ed., McGraw-Hill (2011): Eq. 13-10 to 13-13 (interference), 13-43 to 13-47 (worm forces and efficiency), Table 13-6; Table 14-2 (Lewis Y), Eq. 14-6 (Barth, SI), 14-8, 14-13, 14-14, Table 14-8 (ZE), Eq. 14-27 to 14-29 (AGMA Kv); Examples 14-1, 14-3, 14-4, 13-10, 15-3; Eq. 15-38 (worm friction); Eq. 2-21 (Sut = 3.4 HB). SI 11th ed. (2015): Fig. 14-5, Sc = 2.22 HB + 200 MPa.
- V. B. Bhandari, Design of Machine Elements, 3rd ed., Tata McGraw-Hill: Eq. 17.16-17.17 (beam strength, σb = Sut/3), 17.20-17.22 (Cv), 17.31-17.35 (wear strength, K), 18.24 (helical wear), 19.11-19.14 (bevel factor, face width, bevel wear strength); Examples 17.9, 17.10, 18.4, 19.1, 19.4. The same spur gear forms are in Aliah University's open lecture notes.
- I. B. Celik, U. Ghia, P. J. Roache, C. J. Freitas, "Procedure for estimation and reporting of uncertainty due to discretization in CFD applications", ASME Journal of Fluids Engineering 130 (2008) 078001 (in the journal's editorial policy): Eq. 1-7, Table 1, refinement factor above 1.3.
- Ansys Mechanical 2025 R2 help: Deformation ("calculated relative to the part or assembly world coordinate system"), result coordinate systems, joint probes, Relative Deformation (Custom Results add-on); Ansys Knowledge, stress singularities and mesh refinement (2023).
- Materials: IS 1865:1991 (SG 500/7), IS 1030:1998 (grade 340-570), Ductile Iron Society, Ductile Iron Data for Design Engineers (E 162-170 GPa, ν 0.275), Copper Development Association C90700, Hardin and Beckermann (SFSA 2006, cast steel E 207 GPa), Ansys Granta EduPack data booklet (2021) for a check of the reports' values.
- Skill-Lync, the 697 public gear challenge reports, read 22 Nov 2026. scripts/verify/gear.py reproduces the KHK, Shigley, Bhandari and Celik numbers with its own code (brentq for the involute, a numerical law-of-gearing path of contact, a parametric involute generator, a force balance for the worm) and the page checks itself against it at build time.