Mechanical Engineering Design · Part 3 — Design of Mechanical Elements

Bevel and Worm Gears

Extending the AGMA strength methodology to the harder cases: straight-bevel gear bending and contact stresses with their special factors, the worm-gear allowable-load and thermal-capacity equations, worm friction and efficiency, and Buckingham's classical wear load.

Learning objectives

Recognize when intersecting-shaft bevel gearing or high-ratio worm gearing fits a transmission requirement.
Draw gear free-body diagrams and resolve transmitted, radial, and axial tooth-force components.
Select bending, contact, efficiency, self-locking, wear, and thermal equations for the applicable failure mode.
Separate calculated stress or load from material strength and standards-based allowable values.
Iterate geometry, material, lubrication, mounting, and cooling while checking manufacturability and reliability.
Audit units, force directions, limiting behavior, and the provenance of AGMA or vendor factors.

15–1Bevel Gearing — General

Geometry Chooses the Gear Family

Bevel gears transmit motion between intersecting shaft axes. Straight teeth are economical at moderate speed; spiral teeth engage gradually and generally offer smoother, quieter operation. Hypoid gears serve offset rather than intersecting axes and introduce additional sliding. These families are not interchangeable in a rating calculation: tooth geometry, manufacturing accuracy, lubrication, and the applicable standard determine the factors and allowable loads.

Pitch Cones and Mounting

For a right-angle pair, the pinion and gear pitch angles satisfy \(\tan\gamma=N_P/N_G\) and \(\Gamma=90^\circ-\gamma\). Geometry fixes ratio, but it does not guarantee load sharing. Overhung mounting, shaft slope, bearing clearance, housing distortion, and assembly error can move contact toward a tooth edge. Crowning can tolerate small misalignment, but it cannot repair an inadequate shaft or housing.

Physical meaning

The pitch cones roll together at their apex. Increasing the tooth-count ratio narrows the pinion cone and changes how the same tangential load divides into radial and axial reactions. Those reactions belong on the shaft and bearing free-body diagrams even though only the tangential component transmits useful torque.

15–2Bevel-Gear Stresses and Strengths

Load, Local Stress, and Strength Are Different Quantities

The transmitted load follows from torque at the selected pitch diameter: \(W^t=2T/d_P\). Tooth-root bending stress describes a local response to that load; contact stress describes the Hertzian surface response. Bending strength and contact strength are resistance values obtained from qualified material and heat-treatment data, then modified for life, reliability, and temperature. A nominal torque is therefore not a stress, and a calculated stress is not an allowable strength.

Bevel Bending Stress (U.S. units)

\[ \sigma_t = \frac{W^t\,P_d}{F}\,K_o\,K_v\,\frac{K_s\,K_m}{K_x\,J} \]Eq. (15–3)

Bevel Contact Stress (U.S. units)

\[ \sigma_c = C_p\sqrt{W^t\,K_o\,K_v\,\frac{K_m}{F\,d_P}\,\frac{C_s\,C_{xc}}{I}} \]Eq. (15–1)

Applicability and Failure Modes

These rating forms assume compatible involute bevel geometry, adequate rim support, controlled alignment, and factors taken from the correct edition and unit system of the governing standard. Root fatigue and pitting are distinct modes, so both margins are required. Scuffing, tooth fracture under shock, plastic deformation, noise, shaft deflection, and lubricant starvation require separate checks.

Straight-bevel force components

\[ W^r=W^t\tan\phi\cos\gamma,\qquad W^a=W^t\tan\phi\sin\gamma \]

15–3AGMA Equation Factors

Factors Carry Evidence, Not Decoration

\(K_o\) represents overload duty, \(K_v\) dynamic excitation, \(K_s\) size, and \(K_m\) nonuniform face loading. The geometry factors \(I\) and \(J\), curvature factor \(K_x\), and crowning term \(C_{xc}\) depend on the defined tooth system. Values must be traceable to a named standard, verified software, drawing, or supplier rating. Reusing a spur-gear factor or an unexplained classroom value can create a precise-looking but unauditable result.

Bevel Geometry & Bending Safety Factor

\[ S_F = \frac{s_{at}\,K_L/(K_T\,K_R)}{\sigma_t} \qquad\qquad S_H = \frac{s_{ac}\,C_L\,C_H/(K_T\,C_R)}{\sigma_c} \]Eqs. (15–9, 15–10)

Physical meaning

A factor greater than one usually increases calculated demand or reduces allowable resistance. Before combining factors, inspect what physical uncertainty each one represents and confirm that two factors are not penalizing the same effect twice.

15–4Straight-Bevel Gear Analysis

From Torque to Bearing Loads

Isolate one gear, place \(W^t\) tangent to its pitch cone in the direction opposing motion, and resolve \(W^r\) and \(W^a\) using a declared shaft coordinate. The mating gear receives equal and opposite contact forces, but its radial and axial labels refer to a different shaft axis. Use the complete force system for shaft and bearing reactions before evaluating deflection and contact alignment.

Problem-solving strategy

Recognize whether the target is kinematics, tooth force, stress rating, or component selection.
Set the gear ratio, pitch angles, rotation directions, and positive shaft coordinates.
Convert power and speed to torque, then torque to tangential tooth load at the correct diameter.
Resolve radial and axial components and carry them to shaft-and-bearing free-body diagrams.
For rating, obtain every geometry, life, reliability, and load factor from documented sources.
Check both root bending and surface contact, then iterate geometry and mounting if either fails.

EXAMPLE 1 Resolve a Right-Angle Bevel-Gear Load

A 24-tooth straight-bevel pinion drives a 48-tooth gear on perpendicular shafts. The pinion pitch diameter is \(120\ \text{mm}\), transmitted torque is \(240\ \text{N}\!\cdot\!\text{m}\), and pressure angle is \(20^\circ\). Determine the pinion tangential, radial, and axial tooth-force magnitudes for shaft-bearing analysis.

Recognition strategy

The requested bearing loads identify a bevel-gear force-resolution problem, not a tooth strength calculation. Torque supplies \(W^t\), while tooth ratio supplies the pinion pitch angle needed to split the separating force.

Given information

\(N_P=24\), \(N_G=48\), \(d_P=0.120\ \text{m}\), and \(T_P=240\ \text{N}\!\cdot\!\text{m}\).
\(\phi=20^\circ\); shafts intersect at \(90^\circ\); steady ideal tooth contact is assumed.
Positive axial force is taken from the pinion apex toward its large end; only magnitudes are requested.

Model and interpretation

Isolate the pinion at the mean pitch surface. The tangential component opposes pinion rotation, the radial component points toward the pinion axis under this convention, and the axial component acts along that axis. Housing reactions are excluded from this tooth-contact free body.

Step-by-step solution

1. Find pitch angle. For perpendicular shafts, \(\gamma=\tan^{-1}(N_P/N_G)=\tan^{-1}(24/48)=26.565^\circ\), so the component equations use the pinion cone rather than the gear cone.

2. Convert torque to tangential load. \(W^t=2T_P/d_P=2(240\ \text{N}\!\cdot\!\text{m})/(0.120\ \text{m})=4000\ \text{N}\). This is the useful torque-transmitting component.

3. Resolve the radial component. \(W^r=W^t\tan\phi\cos\gamma=(4000)(\tan20^\circ)(\cos26.565^\circ)=1302\ \text{N}\), directed radially according to the declared pinion free body.

4. Resolve the axial component. \(W^a=W^t\tan\phi\sin\gamma=(4000)(\tan20^\circ)(\sin26.565^\circ)=651\ \text{N}\). Its algebraic sign must be assigned from the actual hand of rotation and mesh orientation.

Verification

Units: Torque divided by pitch radius gives \((\text{N}\!\cdot\!\text{m})/\text{m}=\text{N}\); the trigonometric multipliers are dimensionless, so all three reported components are forces.

Signs and direction: The listed values are magnitudes. Tangential force must oppose the driving rotation, while the axial sign must be transferred to the shaft diagram using the stated positive axis and actual viewing direction.

Plausibility: The separating-force magnitude is \(W^t\tan20^\circ=1456\ \text{N}\), and \(\sqrt{(1302)^2+(651)^2}=1456\ \text{N}\), independently confirming the resolution.

Final result: \(\boxed{W^t=4.00\ \text{kN},\quad W^r=1.30\ \text{kN},\quad W^a=0.651\ \text{kN}}\) as pinion tooth-force magnitudes.

Engineering interpretation

The \(651\ \text{N}\) thrust is not optional simply because torque is tangential. Bearing selection and shaft deflection must include all components; tooth stress, alignment under load, fatigue duty, and mounting tolerances remain separate design checks.

15–5Design of a Straight-Bevel Gear Mesh

Analysis Becomes an Iteration

Begin with ratio, envelope, life, duty, reliability, noise, and manufacturing constraints. Choose standard tooth numbers, module or diametral pitch, face width, quality, material, and heat treatment; then recompute forces, factors, stresses, deflection, and margins. If contact governs, increased diameter, improved alignment, surface hardness, or geometry may help. If bending governs, module, face width, root geometry, and core strength are more direct levers. Face width cannot be increased indefinitely because cone geometry and load distribution limit its benefit.

15–6Worm Gearing — AGMA Equation

Sliding Contact Changes the Design Priority

A worm's screw-like action can provide a large ratio in one mesh, but substantial sliding converts input power into heat. The allowable tangential load relation combines mean gear diameter \(D_m\), effective face width \(F_e\), and empirical material, ratio, and velocity terms. These terms are unit-system and standard specific; they must be sourced rather than inferred from dimensional analysis.

Allowable Tangential Load (worm gear)

\[ (W^t)_{all} = C_s\,D_m^{0.8}\,F_e\,C_m\,C_v \]Eq. (15–28)

Friction Force & Sliding Velocity

\[ W_f = \frac{f\,W^t}{\cos\phi_n\sin\lambda + f\cos\lambda} \qquad\qquad V_s = \frac{\pi\,d_m\,n_W}{12\cos\lambda} \]Eqs. (15–61, 15–62)

Physical meaning

Increasing lead angle generally improves forward efficiency because more contact force acts in the useful circumferential direction. It also reduces the tendency to resist back-driving. Lubricant viscosity, surface finish, material pairing, temperature, and sliding speed all affect friction, so a single coefficient is only valid for its stated operating condition.

15–7Worm-Gear Analysis

Efficiency, Heat, and Back-Driving

When the worm drives, \(\eta\) relates output power to input power and the loss is \(\dot Q_{\mathrm{loss}}=P_{\mathrm{in}}(1-\eta)\). The self-locking inequality is a static equilibrium criterion, not a safety device: vibration, wear, lubricant changes, or reduced static friction can permit motion. Reversing which member drives also changes the applicable force and efficiency model.

Mechanical Efficiency (worm driving)

\[ e_W = \frac{\cos\phi_n - f\tan\lambda}{\cos\phi_n + f\cot\lambda} \]Eq. (15–57)

Self-Locking Condition

\[ f_{stat} \ge \cos\phi_n\,\tan\lambda \]Eq. (15–59)

EXAMPLE 2 Efficiency and Back-Drive Assessment

A steel worm drives a bronze gear with normal pressure angle \(\phi_n=20^\circ\), lead angle \(\lambda=18^\circ\), and operating friction coefficient \(f=0.060\), supplied by a lubricant test at the specified speed and oil temperature. For \(5.00\ \text{kW}\) input, find efficiency, output power, heat generation, and whether a supplied static coefficient \(f_{\mathrm{stat}}=0.11\) predicts self-locking.

Recognition strategy

Specified lead angle and friction identify the worm-driving efficiency model. The power balance gives heat, while back-driving requires the separate static-friction inequality rather than the operating coefficient used for efficiency.

Given information

\(\phi_n=20^\circ\), \(\lambda=18^\circ\), \(f=0.060\), and \(f_{\mathrm{stat}}=0.11\).
\(P_{\mathrm{in}}=5.00\ \text{kW}\); steady worm-driving operation and negligible auxiliary losses are assumed.
Friction values are supplied test data and require confirmation for the production lubricant and temperature.

Model and interpretation

The gearbox is the control volume. Positive power enters through the worm, useful power leaves through the gear, and the remainder is heat. Self-locking is assessed at impending reverse motion with the gear attempting to drive the worm.

Step-by-step solution

1. Evaluate the efficiency numerator. \(\cos20^\circ-(0.060)\tan18^\circ=0.92020\). A positive numerator predicts forward motion without the singular behavior associated with an unsuitable angle-friction combination.

2. Evaluate efficiency. The denominator is \(\cos20^\circ+(0.060)\cot18^\circ=1.12435\), hence \(\eta=0.92020/1.12435=0.8184\), or \(81.84\%\).

3. Close the power balance. \(P_{\mathrm{out}}=\eta P_{\mathrm{in}}=(0.8184)(5.00)=4.092\ \text{kW}\), and \(\dot Q_{\mathrm{loss}}=5.00-4.092=0.908\ \text{kW}\).

4. Test static self-locking. The threshold is \(\cos\phi_n\tan\lambda=(\cos20^\circ)(\tan18^\circ)=0.3053\). Since \(f_{\mathrm{stat}}=0.11<0.3053\), this model does not predict self-locking.

Verification

Units: Efficiency and the friction comparison are dimensionless. Multiplying the dimensionless efficiency by kilowatts returns kilowatts, and input minus output leaves \(0.908\ \text{kW}\) of thermal loss.

Signs and direction: Positive input and output denote forward worm-driven power. Heat loss is positive leaving the mechanical power stream; the failed inequality means reverse motion is not statically blocked by friction alone.

Plausibility: \(0<\eta<1\), output plus loss equals exactly \(5.000\ \text{kW}\) within rounding, and lowering \(f\) in the formula would move efficiency toward one as expected.

Final result: \(\boxed{\eta=81.84\%,\quad P_{\mathrm{out}}=4.092\ \text{kW},\quad \dot Q_{\mathrm{loss}}=0.908\ \text{kW};\ \text{not self-locking}}\).

Engineering interpretation

The housing and lubricant must continuously reject about \(0.91\ \text{kW}\) at this duty. Because self-locking is not predicted—and should never be the sole holding safeguard anyway—a brake or positive stop is required where uncontrolled reverse motion is hazardous.

15–8Designing a Worm-Gear Mesh

Strength and Thermal Capacity Are Parallel Constraints

Select ratio and approximate center distance, then choose a catalog-compatible worm thread count, gear tooth count, diameters, face width, materials, and lubrication method. Compute output tooth load and mechanical rating, then compute heat generation and housing heat rejection at the specified ambient temperature. Change size, ratio, lubrication, fan capacity, or duty cycle and repeat until both constraints and the envelope pass.

Good Worm Pitch-Diameter Proportion

\[ \frac{C^{0.875}}{3} \le d \le \frac{C^{0.875}}{1.7} \]Eq. (15–27)

Worm-set design loop

Translate ratio, output torque, life, duty cycle, ambient temperature, and envelope into requirements.
Select compatible standard geometry and a material-lubricant pair with documented ratings.
Calculate efficiency, actual tooth load, service-adjusted load, and heat generation.
Compare demand independently with tooth capacity and temperature-dependent heat rejection.
Iterate the center distance or cooling method; then verify bearings, shafts, wear, and sealing.

EXAMPLE 3 Select a Worm Reducer by Load and Heat

A continuous conveyor requires \(7.50\ \text{kW}\) at \(90\ \text{rev/min}\). A preliminary worm reducer has efficiency \(0.78\), gear pitch diameter \(320\ \text{mm}\), documented allowable tangential load \(5.20\ \text{kN}\), and housing conductance \(UA=52\ \text{W/K}\). With application factor \(K_o=1.15\), ambient \(35^\circ\text{C}\), and maximum oil temperature \(90^\circ\text{C}\), decide whether it passes. If not, assess a larger catalog unit rated \(6.50\ \text{kN}\) with \(UA=68\ \text{W/K}\) at the same efficiency and pitch diameter.

Recognition strategy

This is a component-selection iteration with two independent limit states. Output torque gives actual tooth load, the supplied application factor gives design load, and the power loss must remain below the housing's supplied steady heat-rejection model.

Given information

\(P_{\mathrm{out}}=7.50\ \text{kW}\), \(n_G=90\ \text{rev/min}\), \(\eta=0.78\), and \(D_m=0.320\ \text{m}\).
Preliminary: \((W^t)_{\mathrm{all}}=5.20\ \text{kN}\), \(UA=52\ \text{W/K}\); larger: \(6.50\ \text{kN}\), \(68\ \text{W/K}\).
\(K_o=1.15\), \(T_\infty=35^\circ\text{C}\), and \(T_{\mathrm{oil,max}}=90^\circ\text{C}\), all supplied design data.

Model and interpretation

Use steady power flow with constant stated efficiency and lumped heat rejection \(\dot Q_{\mathrm{cap}}=UA(T_{\mathrm{oil,max}}-T_\infty)\). The load rating is treated as a catalog allowable, not a calculated material strength; no credit is taken for intermittent cooling because duty is continuous.

Step-by-step solution

1. Convert output power to torque. \(\omega_G=2\pi(90)/60=9.4248\ \text{rad/s}\), so \(T_G=P_{\mathrm{out}}/\omega_G=7500/9.4248=795.8\ \text{N}\!\cdot\!\text{m}\).

2. Form the service-adjusted tooth load. \(W^t=2T_G/D_m=2(795.8)/0.320=4.974\ \text{kN}\), and \(W^t_d=K_oW^t=(1.15)(4.974)=5.720\ \text{kN}\). The preliminary \(5.20\ \text{kN}\) rating fails.

3. Check preliminary thermal capacity. \(P_{\mathrm{in}}=7.50/0.78=9.615\ \text{kW}\), so loss is \(2.115\ \text{kW}\). Capacity is \(52(90-35)=2860\ \text{W}\), which exceeds loss by \(745\ \text{W}\).

4. Assess the larger unit. Its mechanical margin is \(6.50/5.720=1.136\), and thermal capacity is \(68(90-35)=3.740\ \text{kW}\), giving \(3.740-2.115=1.625\ \text{kW}\) thermal margin; both stated checks pass.

Verification

Units: Watts divided by radians per second produce \(\text{N}\!\cdot\!\text{m}\) because radians are dimensionless; torque divided by metres gives newtons, and \((\text{W/K})(\text{K})\) gives watts of heat capacity.

Signs and direction: Output power and torque are positive in the driven direction. Heat generation is positive because input exceeds output, and positive capacity margins mean the listed limit exceeds demand.

Plausibility: The \(78\%\) efficiency implies a \(22\%\) input loss: \(9.615-7.500=2.115\ \text{kW}\). The actual load is just below the first rating, but applying the explicitly required duty factor correctly exposes failure.

Final result: \(\boxed{\text{Reject preliminary unit; select larger unit with load margin }1.136 \text{ and thermal margin }1.625\ \text{kW}}\).

Engineering interpretation

The larger catalog unit passes only the supplied simplified checks. Final release still requires manufacturer confirmation of rating life, lubricant and ambient range, starting torque, shaft and bearing loads, seals, noise, cooling airflow, tolerances, and the applicable AGMA or ISO rating edition.

15–9Buckingham Wear Load

A Screening Relation, Not Automatic Certification

The Buckingham relation packages material-pair wear behavior into \(K_w\). It can screen a candidate when its units, material condition, lubrication basis, and source are known. It does not capture modern reliability targets, transient overload, temperature-dependent lubrication, or detailed contact geometry and should not be mixed silently with AGMA factors.

Buckingham Allowable Wear Load

\[ (W^t_G)_{all} = K_w\,d_G\,F_e \]Eq. (15–63)

Chapter recap

Bevel pitch-cone geometry establishes ratio and the radial and axial forces carried by shafts and bearings.
Bending stress, contact stress, and allowable strengths are separate quantities with separate failure modes.
Worm design couples mechanical load capacity to efficiency, wear, lubrication, and heat rejection.
Document every empirical factor and iterate until geometry, strength, thermal, manufacturing, and reliability constraints pass.

Before moving on: you should be able to draw either mesh free body, select the relevant load or efficiency relation, audit factor provenance, and explain which additional standard, vendor, fatigue, thermal, and test checks remain.