In our work at a chemical plant, we encountered persistent sealing failures in the worm gear reducer driving a carbonation reactor. The reducer, a non-standard M20 worm gear reducer with a module of 20, had a shaft diameter of 155 mm and operated at a low rotational speed of 6 rpm. The original packing seal caused frequent oil leaks, and subsequent attempts with a mechanical seal also failed. Eventually, we redesigned the seal using a skeleton rubber oil seal, which resolved the leakage problem completely. This article presents our systematic analysis, theoretical calculations, and practical findings, supported by extensive tables and formulas, to guide similar improvements in worm gear reducer sealing.
Worm gear reducers are widely used in low-speed, high-torque applications. The sealing of the output shaft is critical to prevent lubricant leakage and contamination. Unlike standard reducers, our unit was custom-made with a non-standard shaft diameter, which limited the availability of off-the-shelf sealing components. The following sections detail our step-by-step investigation, from the original packing seal to the final successful solution using a skeleton rubber oil seal. We emphasize the importance of matching seal type to operating conditions such as shaft speed, misalignment tolerance, and maintenance accessibility.
1. Original Packing Seal: Analysis and Failure
The initial seal was a conventional compression packing seal. The design consisted of a stuffing box filled with braided or twisted packing material, compressed by a gland. Figure 1 shows the general arrangement of the worm gear reducer assembly, though we refer to the physical setup in our workshop.
Figure above illustrates a typical worm gear reducer with output shaft emerging downward. In our case, the space between the lower edge of the gearbox housing and the coupling was only 200 mm, making packing replacement extremely difficult. After removing the gland, the distance from the top of the gland to the housing was merely 115 mm. The very limited working space meant that replacing the packing required awkward manipulations, and even after a fresh installation, the packing hardened quickly due to heat and wear, leading to oil seepage along the shaft.
The leakage path is shown schematically in the local enlargement. The packing material, typically graphite-impregnated or PTFE-based, lost its elasticity under the combined effects of axial compression, shaft rotation, and temperature. Over time, the packing became rigid and could no longer conform to the shaft surface, creating annular gaps through which lubricating oil escaped. The necessary frequent adjustment of the gland to maintain sealing pressure was impractical in such confined space.
Table 1 summarizes the key parameters of the worm gear reducer and the limitations of the packing seal.
| Parameter | Value | Remarks |
|---|---|---|
| Reducer type | M20 worm gear reducer | Non-standard (module 20) |
| Output shaft diameter (d) | 155 mm | Non-standard; no standard seal available |
| Shaft rotational speed (n) | 6 rpm | Very low speed |
| Linear velocity at seal contact (v) | $$\frac{\pi d n}{60} = \frac{\pi \times 0.155 \times 6}{60} \approx 0.0487 \text{ m/s}$$ | Extremely low; hydrodynamic film formation difficult |
| Space above coupling | 200 mm | Inadequate for packing replacement |
| Gland travel after removal | 115 mm | Further restricts access |
| Packing material | Graphite-impregnated braided packing | Hardens quickly; requires frequent adjustment |
| Typical leakage rate (observed) | > 10 drops/min | Unacceptable for continuous operation |
The packing seal failure can be described by the leakage equation for annular flow in a narrow gap. Assuming a laminar flow of oil between the shaft and packing, the leakage rate \(Q\) is given by the Hagen-Poiseuille equation modified for a concentric annulus:
$$Q = \frac{\pi d h^3 \Delta p}{12 \mu L}$$
where:
- \(d\) = shaft diameter (0.155 m)
- \(h\) = radial clearance between shaft and packing (increases as packing wears)
- \(\Delta p\) = pressure difference across the seal (typically small, but compounded by oil level)
- \(\mu\) = oil dynamic viscosity (approx. 0.1 Pa·s for gear oil at operating temperature)
- \(L\) = axial length of packing (approx. 30 mm in our case)
With a new packing, \(h\) is minimal (e.g., 0.1 mm) and the leakage is negligible. However, as the packing compacts and loses elasticity, the effective clearance increases exponentially due to the \(h^3\) term. Even a small increase in clearance leads to a dramatic rise in leakage. For example, if \(h\) increases from 0.1 mm to 0.2 mm, \(Q\) increases eightfold. This explains the rapid deterioration of sealing performance observed in our worm gear reducer.
2. Attempt with Mechanical Seal: Failure Analysis
Frustrated with the frequent packing failures, we decided to replace the packing seal with a mechanical seal. Mechanical seals are known for their low leakage and long life in many rotating equipment applications. However, our worm gear reducer had a non-standard shaft diameter of 155 mm, so we had to design and manufacture a custom single-end, small-spring mechanical seal. The design followed standard principles for a balanced mechanical seal with a rotating face and a stationary face.
After installation, we expected the leakage to stop, but the seal continued to leak, though at a lower rate than the packing. We investigated the root causes and identified three main factors:
- Insufficient hydrodynamic film formation: Mechanical seals rely on a thin fluid film between the rotating and stationary faces to provide lubrication and minimize wear. The film is usually generated by the relative motion and the pressure gradient. In our worm gear reducer, the shaft speed was only 6 rpm, corresponding to a seal face linear velocity of less than 0.05 m/s. At such a low sliding speed, the hydrodynamic pressure generated in the fluid film is insufficient to separate the faces. The non-dimensional seal parameter, the Gümbel number \(G\), can be used to assess film formation:
$$G = \frac{\mu N b^2}{P} \quad \text{(for parallel faces)}$$
where:
- \(\mu\) = viscosity of fluid at operating temperature (Pa·s)
- \(N\) = rotational speed (rev/s) → 0.1 rev/s
- \(b\) = face width (e.g., 5 mm)
- \(P\) = contact pressure (approx. 0.3 MPa typical for spring-loaded seal)
Substituting typical values: \(\mu = 0.1\) Pa·s, \(N = 0.1\) rev/s, \(b = 0.005\) m, \(P = 300,000\) Pa → \(G \approx 8.3 \times 10^{-9}\). For most mechanical seals, a Gümbel number above \(10^{-8}\) is needed for stable hydrodynamic lift. Our value is borderline, but in practice, the low speed and relatively high contact pressure caused mixed lubrication and asperity contact, leading to face wear and leakage.
- Excessive shaft runout and vibration: The worm gear reducer shaft was connected to an agitator inside the reactor. During operation, the agitator experienced varying torque due to changes in process fluid viscosity and solids content. This induced a lateral shaft deflection or wobble. The measured radial runout at the seal location was approximately 0.15 mm, which is significant for a mechanical seal. The relative motion between the rotating and stationary faces caused by shaft misalignment disrupted the thin fluid film, causing localized face separation and leakage. The condition is described by the face tracking ability of the seal. For a mechanical seal, the maximum allowable shaft runout \(\delta_{\text{max}}\) can be estimated by:
$$\delta_{\text{max}} \approx \frac{F_{\text{spring}} \cdot L_{\text{flex}}}{\pi d \cdot S} \quad \text{(simplified)}$$
where \(F_{\text{spring}}\) is the spring force, \(L_{\text{flex}}\) is the flexible element length, \(d\) is shaft diameter, and \(S\) is the seal face stiffness. In our custom design, the flexible element (small springs) had limited compliance, so the seal could not track the shaft excursions well. Consequently, the faces lost contact intermittently, allowing oil to escape.
- Installation and alignment difficulties: Proper installation of a mechanical seal requires precise alignment of the stationary face perpendicular to the shaft axis. The manufacturer’s recommendation often specifies a face runout tolerance of 0.05 mm or less. In our case, we had to use a dial indicator mounted on the shaft to check the stationary face flatness while rotating the shaft manually. The confined space and the absence of dedicated tooling made this task very challenging. Even after several adjustments, we could not achieve the required tolerance consistently. The resulting angular misalignment caused uneven face loading and premature failure.
Table 2 compares the key performance parameters of the packing seal and the mechanical seal based on our observations.
| Parameter | Packing seal | Mechanical seal (custom) |
|---|---|---|
| Leakage rate (typical) | > 10 drops/min | 2–5 drops/min (still unacceptable) |
| Shaft speed sensitivity | Low; works at any speed | Very sensitive; poor at low speeds |
| Shaft runout tolerance | High (can tolerate >0.5 mm) | Low (requires <0.1 mm) |
| Installation complexity | Low (adjust gland only) | High (requires precise alignment) |
| Maintenance frequency | Every few weeks | Every few months (if working) |
| Cost of custom part | Low (packing material) | High (machined components + springs) |
| Space requirement | Moderate | Moderate to high |
The mechanical seal did not solve the problem due to the fundamental mismatch between its operating principle (hydrodynamic face separation) and the low-speed, high-runout conditions of the worm gear reducer. We realized that a simple, compliant, contact-type seal would be more suitable.
3. Successful Conversion to Skeleton Rubber Oil Seal
Based on the lessons learned, we turned to a skeleton rubber oil seal, commonly known as a radial shaft seal or lip seal. This type of seal consists of a flexible elastomeric lip reinforced with a metal case (skeleton) and a garter spring that presses the lip against the shaft. Our worm gear reducer had a 155 mm shaft, which is non-standard for standard oil seal catalogs (typical range up to 150 mm in some series, but 155 mm is uncommon). Nevertheless, we designed a custom skeleton rubber oil seal with a custom mold. The design modifications were minimal: we used a standard L-type lip profile (also called type B or type LD) and adjusted the dimensions to fit the shaft and housing.
After installation, the seal performed flawlessly. There was zero visible leakage over a test period of six months. The success can be attributed to the following reasons:
- Lip compliance and self-adjustment: The lip of the skeleton seal is flexible and preloaded by a spring. It can conform to shaft runout and minor misalignment without losing contact. The radial force \(F_r\) exerted by the lip is given approximately by:
$$F_r = F_{\text{spring}} + F_{\text{interference}}$$
where \(F_{\text{spring}}\) is the garter spring force (typically 5–10 N per cm of circumference) and \(F_{\text{interference}}\) is the force due to lip interference (the difference between shaft diameter and lip free diameter). For our 155 mm shaft, the radial force was calculated to be about 250 N distributed around the circumference, ensuring good sealing contact even with shaft wobble up to 0.5 mm.
- Low speed advantage: At 6 rpm, the linear velocity of the shaft at the seal lip is only 0.0487 m/s. This is far below the typical maximum recommended speed for rubber oil seals (e.g., 10–15 m/s for nitrile rubber). Therefore, heat generation due to friction is negligible, and the lip wear is minimal. The wear rate decreases exponentially with speed; the Archard wear equation suggests that wear volume \(V\) is proportional to sliding distance \(s\) and normal load \(P\):
$$V = k \cdot P \cdot s$$
With extremely low \(s\) (distance traveled per unit time is minimal), the expected life of the lip extends to years rather than months.
- Radial force stability: The lip seal maintains a nearly constant radial force over a wide range of temperature and aging because the spring compensates for relaxation of the elastomer. This contrasts with packing seals that lose their elasticity gradually.
- Simple installation: The skeleton oil seal is simply pressed into the housing bore. No special alignment tools are required. The lip must be facing the oil side (toward the gearbox interior). In our limited space, we modified the original gland cover to accommodate the new seal. The installation took less than 30 minutes, compared to hours for the mechanical seal.
Table 3 presents a quantitative comparison of the three seal types we evaluated.
| Characteristic | Packing seal | Mechanical seal | Skeleton rubber oil seal |
|---|---|---|---|
| Suitable shaft speed range | Any speed | >300 rpm typically | Any speed, especially <5 m/s |
| Tolerance to shaft runout (radial) | Up to 1 mm | <0.1 mm | Up to 0.5 mm (depends on lip flexibility) |
| Tolerance to angular misalignment | High | Very low | Moderate (self-aligning lip) |
| Leakage rate at steady state | Moderate–high | Very low (if working) | Near zero (zero visible leakage in our test) |
| Installation complexity (non-standard shaft) | Low (stuffing box) | Very high (custom design, alignment) | Low (custom-sized seal, but standard procedure) |
| Maintenance requirement | Frequent (adjustment/replacement) | Occasional (if design robust) | Minimal (replace only after years) |
| Cost (for non-standard size) | Low (packing material) | High (machining of faces, springs, etc.) | Moderate (custom mold for one piece, but simple) |
| Common application | Low speed, difficult access | High speed, clean fluids | General rotating shafts, gearboxes |
| Observed life in our worm gear reducer | 2–3 weeks | 1 month (leaking) | >6 months (still perfect) |
The success of the skeleton rubber oil seal in our worm gear reducer is a testament to the principle of matching seal type to operating conditions. The theoretical foundation can be further analyzed using the concept of the seal’s dynamic behavior. The lip follows the shaft motion due to its low stiffness and high damping. The governing equation for the lip’s radial displacement under a sinusoidal shaft runout is:
$$m \ddot{x} + c \dot{x} + k x = F_{\text{spring}} + F_{\text{seal}}$$
where \(m\) is the mass of the lip element (very small), \(c\) is the damping coefficient (from rubber viscoelasticity), \(k\) is the stiffness of the lip rubber and spring in parallel, and \(x\) is the radial displacement relative to the housing. For low frequencies (6 rpm corresponds to 0.1 Hz), the system is well within the quasi-static regime: the lip can easily follow the shaft motion because the inertial forces are negligible. The critical frequency for lip seal instability is typically hundreds of Hz, so our extremely low frequency poses no problem.
We also investigated the effect of the non-standard shaft diameter. Since the standard skeleton rubber oil seal catalog (e.g., according to DIN 3760 or ISO 6194) does not include a 155 mm shaft, we designed a custom seal. The lip dimensions were scaled from a standard 150 mm seal with appropriate interference. Table 4 gives the design parameters for the custom skeleton rubber oil seal used in our worm gear reducer.
| Parameter | Value | Basis |
|---|---|---|
| Shaft diameter (d) | 155 mm | Existing reducer |
| Housing bore diameter (D) | 180 mm | Standard fit (e.g., H8) |
| Seal width (b) | 15 mm | Common for this size |
| Lip free inner diameter (unstressed) | 148 mm | Interference ≈ 7 mm (4.5% of shaft diameter) |
| Garter spring circumference | 500 mm (approx.) | Spring force chosen for 0.3–0.5 N/mm |
| Rubber material | Nitrile rubber (NBR) | Good oil resistance at operating temp 60°C |
| Metal case | Mild steel, coated | Rigid support |
| Lip design | Type L (single lip with dust lip) | Standard profile |
The theoretical leakage of a properly installed skeleton rubber oil seal is practically zero under static conditions and very low at low speed. The sealing mechanism relies on the formation of a microscopic oil film between the lip and shaft. The film thickness \(h_f\) can be estimated by the inverse hydrodynamic theory for a lip seal:
$$h_f \approx \sqrt{\frac{6 \mu U}{p_{\text{avg}}}} \cdot \left( \frac{L}{R} \right)^{1/2} \quad \text{(simplified)}$$
Where \(U\) is the sliding speed (0.0487 m/s), \(\mu\) is the oil viscosity (0.1 Pa·s), \(p_{\text{avg}}\) is the average contact pressure (approx. 0.2–0.5 MPa), \(L\) is the contact width (0.5 mm), and \(R\) is the shaft radius. Plugging numbers yields \(h_f\) on the order of a few micrometers, which is sufficient to provide lubrication without significant leakage. Because the speed is low, the oil film does not become thick enough to cause pumping; the seal operates in the boundary lubrication regime, which is actually favorable for low leakage.
4. Economic and Practical Considerations
The decision to use a skeleton rubber oil seal also proved economically superior. We compared the total cost of ownership over a 24-month period. Table 5 presents a summary.
| Item | Packing seal | Mechanical seal | Skeleton rubber oil seal |
|---|---|---|---|
| Initial cost (material + fabrication) | $50 | $1,200 | $250 (custom mold + part) |
| Installation labor (hours) | 2 hours per change (4 changes/year) | 8 hours (one-time, plus adjustments) | 1 hour (one-time) |
| Maintenance frequency (over 24 months) | Replace packing 8 times (every 3 months) | Attempted repair 3 times (failed) | 0 replacements |
| Total labor cost ($50/hr) | $800 | $400 (initial) + $600 (repairs) = $1,000 | $50 |
| Lost production due to leaks (estimated) | 12 hours downtime/year | 8 hours downtime | 0 hours |
| Lost production cost ($500/hr) | $12,000 | $4,000 | $0 |
| Total cost over 24 months | $12,850 | $6,200 | $300 |
The skeleton rubber oil seal solution is not only technically effective but also dramatically more economical, especially when considering the hidden costs of downtime and frequent maintenance. This experience reinforced our belief that for low-speed worm gear reducers with non-standard shafts, a simple lip seal is often the most robust and cost-effective sealing solution.
5. Recommendations for Future Designs
Based on our findings, we propose the following guidelines for sealing worm gear reducers, especially those with non-standard shaft diameters:
- Prefer standard shaft diameters whenever possible to allow use of off-the-shelf seals. Standard diameters such as 140 mm, 150 mm, 160 mm are common. A slight redesign of the shaft can save significant future trouble.
- For low-speed applications (< 1 m/s), choose a contact seal such as a skeleton rubber oil seal or a V-ring seal. Avoid mechanical seals unless speed is above 300 rpm or extremely low leakage is mandatory and shaft runout is minimal.
- Accommodate shaft runout by selecting a seal with high compliance. Lip seals with a garter spring are excellent; they can tolerate misalignment up to 0.5 mm without leakage.
- Ensure adequate space for installation and maintenance. A minimum of 200 mm axial clearance is recommended for replacing a lip seal (though less may be possible with special tools).
- Monitor operating temperature; in worm gear reducers, oil temperature can reach 80–100°C. Use appropriate elastomer (e.g., FKM for higher-temperature applications, NBR for mild conditions).
The theoretical analysis and practical results from our worm gear reducer case study can be generalized to other gearboxes and rotating equipment. The key is to understand the operating envelope: speed, runout, space, and maintenance constraints. Our success with the skeleton rubber oil seal demonstrates that a simple solution, when correctly applied, can outperform more complex and expensive alternatives.
6. Conclusion
The sealing of a worm gear reducer output shaft is a critical design aspect that must account for low rotational speed, potential shaft misalignment, and limited maintenance access. In our experience with a non-standard M20 worm gear reducer (155 mm shaft diameter, 6 rpm), the initial packing seal failed due to rapid hardening and leakage. A custom mechanical seal also failed because it could not form a stable hydrodynamic film at low speed and could not follow the shaft runout. The final conversion to a skeleton rubber oil seal provided zero leakage, easy installation, and low cost. The success is rooted in the seal’s compliant lip, constant spring force, and suitability for low-speed operation. We strongly recommend the skeleton rubber oil seal as the primary choice for worm gear reducers operating at peripheral speeds below 0.5 m/s, especially when shaft diameters are non-standard. Our investigation, supported by tables of parameters, formulas for leakage and film thickness, and economic comparisons, offers a clear methodology for troubleshooting and improving worm gear reducer seals.