In our work on permanent magnet DC gearmotors for automatic doors and barrier gates, we encountered significant limitations in traditional worm gear drives derived from automotive windshield wiper motors. These conventional worm gears suffered from low efficiency, poor load capacity, and excessive wear. Over several design iterations, we systematically improved the worm gear pair by analyzing forces, friction losses, and geometric parameters. In this article, we share our design methodology, key formulas, and results from our experiments.
Force and Loss Analysis in Worm Gears
We began by examining the forces acting on the worm and worm wheel. For a worm gear pair, the tangential force on the worm Ft1 equals the axial force on the worm wheel Fx2 (with opposite sign). The fundamental relationships are given below. Note that all torques are in N·mm, and forces in N.
| Force component | Expression | Description |
|---|---|---|
| Worm tangential force | $$F_{t1} = -F_{x2} = \frac{200\,T_1}{d_1}$$ | Opposes applied torque \(T_1\) |
| Worm axial force | $$F_{x1} = -F_{t2} = -\frac{200\,T_2}{d_2 + 2x_2 m}$$ | Produced by wheel torque \(T_2\) |
| Worm radial force | $$F_{r1} = -F_{r2} = -F_{t2} \tan \alpha_x$$ | Points toward centers |
| Normal force | $$F_n = \frac{F_{x1}}{\cos\gamma \cos\alpha_n} \approx -\frac{200\,T_2}{d_2 \cos\gamma \cos\alpha_x}$$ | Perpendicular to tooth flank |
From these forces we derived the sliding velocity \(V_S\) and friction loss power. The efficiency \(\eta\) of a worm gear (worm driving) is:
$$ \eta = \frac{\tan\gamma}{\tan(\gamma + \rho_v)} $$
where \(\gamma\) is the lead angle and \(\rho_v\) is the equivalent friction angle. The sliding velocity is:
$$ V_S = \frac{V_1}{\cos\gamma} = \frac{d_1 n_1}{19090 \cos\gamma} = \frac{m n_1}{19090 \sin\gamma} $$
Here \(m\) is the axial module in mm, \(n_1\) the worm speed in rpm. The friction force along the helix is:
$$ F = \mu F_n = \frac{200 \mu T_2}{d_2 \cos\gamma \cos\alpha_x} $$
And the friction power loss becomes:
$$ P = F V_S = \frac{200 \mu T_2 m n_1}{19090\,d_2 \cos\gamma \sin\gamma \cos\alpha_x} = \frac{400 \mu T_2 m n_1}{19090\,d_2 \sin 2\gamma \cos\alpha_x} $$
We also considered worm shaft deflection. For a worm supported at both ends, the deflection at the meshing point is:
$$ y = \frac{\sqrt{F_{t1}^2 + F_{r1}^2}\,L^3}{48 E I} $$
For a cantilever support (one end fixed):
$$ y = \frac{\sqrt{F_{t1}^2 + F_{r1}^2}\,L^3}{24 E I} $$
These equations clearly show that to reduce friction losses we need to: (1) increase lead angle \(\gamma\), (2) decrease pressure angle \(\alpha_n\), (3) reduce the coefficient of friction \(\mu\) by improving surface hardness, and (4) increase worm shaft stiffness to maintain proper meshing.
Optimizing Geometric Parameters
Based on the force and loss analysis, we decided to modify the worm geometry. Traditionally, worm gear designs use a characteristic factor \(q = d_1 / m\) of about 8–12. We reduced \(q\) to make the worm diameter smaller and increase the lead angle for a given number of starts. However, a smaller worm diameter reduces bending strength and stiffness. We compensated by using a reduced tooth height (short tooth profile) and sometimes a dual-module calculation for worm and wheel geometry. The key parameters we altered are summarized below.
| Parameter | Conventional | Improved | Benefit |
|---|---|---|---|
| Pressure angle \(\alpha_n\) | 20° | 8° | Lower radial force, higher stiffness |
| Lead angle \(\gamma\) | 5°–10° | ~9.78° | Higher efficiency |
| Characteristic factor \(q\) | 10 | 6–8 | Smaller worm, larger lead angle |
| Tooth height | 2.25 \(m\) | 1.6 \(m\) | Stronger root, smaller deflection |
| Shaft support | Cantilever | Both ends | Deflection halved |
We also increased the worm surface hardness by nitriding to achieve a case depth of 0.3–0.5 mm and hardness above 700 HV. This reduced the friction coefficient \(\mu\) from about 0.08 to 0.04 under typical lubricated conditions.
Application Example: Improved Worm Gear Design
We designed a worm gear pair for a DC motor rated at 24 V, 150 W, with output speed of 30 rpm. The worm had 1 start, axial module \(m_x = 1.25\) mm, lead angle \(\gamma = 9.7824^\circ\) (right hand), and axial pressure angle \(\alpha_n = 8^\circ\). The worm wheel had 40 teeth. The worm diameter was reduced to \(d_1 = q \cdot m = 8 \times 1.25 = 10\) mm (versus conventional 14 mm). The tooth thickness was adjusted based on wear tests to achieve optimal contact pattern.
The resulting efficiency measured on a dynamometer was 72%, compared with 66% for the conventional design – an improvement of 6 percentage points. The worm shaft deflection under rated load was only 0.02 mm, well within acceptable limits. The gearbox temperature rise dropped by 8°C due to reduced friction heating.
Figure 1 below shows the tooth profile of the improved worm. The reduced pressure angle is clearly visible compared to a standard 20° involute.
Blanking Blank Calculation for Forming Processes
While focusing on worm gears, we also encountered a related forming process for the gear blank. In our production line, we use a drum-shaped blank (before hobbling) to reduce material waste. The blank dimensions are calculated using the formula for a drum shape (see sketch in our design guide). The key dimension \(a\) (the flat width before forming) is given by:
$$ a = 2\sqrt{(R+t)^2 – \left(\frac{H}{2}\right)^2} – 2t $$
where \(R\) is the drum radius, \(H\) the workpiece height, and \(t\) the blank thickness. This formula ensures that after forming the drum, the final gear blank has the correct cylindrical shape with minimal flash. We adjusted the gap between inner and outer dies to exactly one material thickness, which is controlled by varying the pulley diameter and guide plate thickness. For flexibility, we made the wedge adjustable.
| Parameter | Symbol | Value (mm) |
|---|---|---|
| Drum radius | \(R\) | 25 |
| Workpiece height | \(H\) | 12 |
| Blank thickness | \(t\) | 3 |
| Computed flat width | \(a\) | 46.8 |
Further Design Considerations for Worm Gears
We also investigated the effect of worm gear ratio on efficiency. Our tests showed that for a given center distance, a larger lead angle (achieved by using multi-start worms) increases efficiency significantly, but reduces the self-locking capability. For applications requiring self-locking (e.g., elevator doors), we limited the lead angle to below 6°. For non-self-locking applications (e.g., barrier gates), we used lead angles up to 15°.
Lubrication played a critical role. We switched from mineral oil to synthetic polyalphaolefin (PAO) grease with molybdenum disulfide additive. This reduced the friction coefficient by another 10–15% and extended the life of the worm gear pair beyond 2000 hours of continuous operation.
| Lubricant type | Friction coefficient \(\mu\) | Efficiency \(\eta\) (%) |
|---|---|---|
| Mineral oil ISO VG 150 | 0.07 | 66 |
| PAO grease with MoS₂ | 0.03 | 74 |
| Ester-based grease | 0.04 | 71 |
Stiffness and Deflection Analysis
We performed finite element analysis to verify the worm shaft deflection. The analytical formula gave good agreement with FEA results. For the improved design with both ends supported, the maximum deflection at the meshing point was 0.018 mm under rated load. The bending stress was below 50 MPa, well within the endurance limit of the 20MnCr5 case-hardened steel we used.
For completeness, we derived the stiffness requirement. The allowable deflection for proper contact is usually less than 0.05 mm. Our design achieved this by: (a) using a larger root diameter (by reducing tooth height), (b) supporting both ends with deep groove ball bearings, and (c) optimizing the helix angle to balance axial forces.
Experimental Validation
We built a test rig to measure efficiency, temperature, and wear over 1000 hours. The improved worm gears showed an average efficiency of 73% compared to 65% for the baseline. Temperature rise at the gearbox housing was 35°C versus 48°C. Visual inspection after 500 hours showed minimal pitting on the worm wheel (bronze) and no measurable wear on the hardened worm.
| Parameter | Conventional design | Improved design | Improvement |
|---|---|---|---|
| Efficiency at rated load | 65% | 73% | +8% |
| Temperature rise (ΔT) | 48°C | 35°C | -13°C |
| Worm deflection (μm) | 45 | 18 | 60% less |
| Noise level at 1m (dB) | 62 | 55 | Quieter |
Conclusion
Through systematic analysis of forces, friction, stiffness, and geometry, we successfully improved the efficiency of worm gears used in DC gearmotors. By reducing the pressure angle, increasing the lead angle, optimizing the worm diameter, and using both-end support, we achieved a 6–8% improvement in efficiency while maintaining load capacity and reliability. The design changes are straightforward to implement in existing gearbox housings, making them suitable for retrofit applications. Our work demonstrates that even mature technologies like worm gears can be significantly enhanced by careful application of fundamental mechanics.
We continue to explore further improvements, such as using asymmetric tooth profiles and advanced surface coatings. The formulas and tables presented here serve as a practical guide for engineers aiming to optimize worm gear drives for low-power DC motors.