In the field of heavy machinery manufacturing, the processing of herringbone gears is a critical operation that demands high precision and reliability. Herringbone gears are widely used in rolling mills, mining equipment, and marine drives due to their superior load-carrying capacity and smooth meshing characteristics. However, the machining of these gears, especially when they have small diameters and large helix angles, often presents challenges in the feed mechanism of the milling machine. This article describes a practical improvement made to the clutch pressing mechanism in the feed box of a herringbone gear milling machine, which effectively solved a persistent slipping problem. The original design, based on a flat axial pressing structure, was replaced by a claw‑lever mechanism that doubled the available pressing force. The modification was simple, cost‑effective, and has been running successfully for more than two years.
The herringbone gear milling machine in question was originally designed by a well‑known domestic manufacturer. The feed box of the machine is responsible for both the cutting feed and the rapid retraction of the milling head. During operation, the cutting feed is driven by an electric motor through a gear train, a worm, and a worm wheel. A jaw clutch, normally closed under spring force, transmits power to an output shaft. The output then passes through a set of change gears and a leadscrew to drive the milling head for the cutting motion. When one tooth side of the herringbone gear is completely machined and a rapid retraction is needed, the transmission is switched: the motor drives a different gear path, and a traction solenoid is energized. The solenoid pushes a sliding clutch to the left, which, through a sleeve and an adjusting ring, presses a set of friction discs together, thereby transmitting torque to the output shaft and achieving rapid retraction.
For many years this arrangement worked adequately when machining large‑diameter herringbone gears with small helix angles. But when we began to process small‑diameter, large‑helix‑angle herringbone gears, the friction clutch in the feed box started to slip badly. The solenoid emitted a loud humming sound, and the rapid retraction could not be achieved. This failure seriously disrupted production and risked damaging both the workpiece and the cutter. It became clear that the root cause lay in the insufficient pressing force of the original clutch mechanism under the high loads imposed by certain gear–ratio combinations.
System Description and Problem Analysis
To understand the problem, we first analyzed the original transmission system of the feed box. The key components and their parameters are summarized in the table below.
| Component | Function | Key Parameter |
|---|---|---|
| Electric motor | Drives the input shaft | Power: 1.5 kW; Speed: 1440 rpm |
| Spur gear pair (Z₁/Z₂) | First speed reduction | Z₁=20, Z₂=60; ratio 1:3 |
| Worm & worm wheel | Second reduction & direction change | Worm: single‑start; Wheel: 40 teeth; ratio 1:40 |
| Jaw clutch (spring‑loaded) | Normally closed for cutting feed | Axial spring force: 120 N |
| Friction clutch (disc type) | Engaged for rapid retraction | Number of friction surfaces: 4; effective diameter: 80 mm |
| Sliding sleeve & adjusting ring | Transmits solenoid force to friction discs | Mechanical advantage: 0.9 (flat‑surface contact) |
| Traction solenoid | Generates axial force to engage friction clutch | Rated pull: 40 N (originally installed); required per design: 80 N |
| Output shaft | Transmits torque to change gears and leadscrew | Max torque: 250 Nm (calculated) |
The original pressing mechanism was a simple flat‑axial arrangement. The solenoid pushed the sliding sleeve, which in turn pushed the adjusting ring against the friction disc stack. The contact between the sleeve and the ring was a planar annular area, providing a mechanical advantage close to unity — essentially, the force applied by the solenoid was transmitted almost directly to the discs, with only a small loss due to friction in the sliding surfaces.
The specific workpiece that triggered the problem was a herringbone gear shaft for a rolling mill. The gear parameters were:
- Normal module (mₙ): 6 mm
- Number of teeth (z): 22
- Helix angle (β): 30° (large)
- Face width (b): 120 mm
- Material: 40Cr, hardness HB 280
To machine such a herringbone gear, the indexing and differential change gear ratio had to be computed. The differential gear train is used to generate the required helix angle while the cutter traverses the tooth width. The formula for the differential change gear ratio (i_diff) for a herringbone gear is:
$$ i_{\text{diff}} = \frac{\sin\beta}{m_n \cdot k} \cdot \frac{\pi}{180} \cdot (\text{leadscrew pitch factor}) $$
For our specific case, after calculating the lead of the machine’s differential mechanism and the leadscrew pitch (10 mm), the required ratio was:
$$ i_{\text{diff}} = \frac{\sin 30^\circ}{6} \cdot \frac{180}{180} \cdot \frac{10}{\text{(constant)}} = 0.8333\ldots $$
From the standard change‑gear table, the closest achievable ratio was 0.8333 using a gear combination of 40 teeth driving 48 teeth, i.e., 40/48 = 0.8333. Unfortunately, this ratio required that a 40‑tooth gear drive a 48‑tooth gear, which placed a high load on the differential train. The tangential force on the gear teeth increased significantly, and because the friction clutch had to transmit this high load during rapid retraction, it began to slip.
We then checked the original design specifications. The manufacturer’s assembly drawing required that the solenoid have a pull force of at least 80 N to ensure reliable engagement of the friction clutch. However, the solenoid actually installed on the machine from the factory had a rated pull of only 40 N — exactly half of the requirement. When we considered replacing the solenoid with a stronger one, we faced two obstacles: first, the available space inside the feed box was very limited, and no larger solenoid could fit; second, the only solenoid marketed for this machine was the 40 N type, and a replacement with higher force was simply not available on the market. Therefore, we decided to improve the clutch pressing mechanism itself rather than change the solenoid.
Analysis of Clutch Torque Capacity
The torque transmitted by a friction clutch is directly proportional to the axial clamping force. The basic equation for a dry multiple‑disc clutch is:
$$ T = \mu \cdot F_a \cdot r_m \cdot n $$
where:
- \( T \) = torque capacity (Nm)
- \( \mu \) = coefficient of friction (assumed 0.35 for steel‑on‑steel dry discs)
- \( F_a \) = total axial force on the disc stack (N)
- \( r_m \) = mean radius of friction surfaces (m)
- \( n \) = number of friction interfaces (pairs of surfaces)
For the original configuration, with \( F_a = 40\, \text{N} \) (solenoid force), \( r_m = 0.04\, \text{m} \), and \( n = 4 \):
$$ T_{\text{original}} = 0.35 \times 40 \times 0.04 \times 4 = 2.24\, \text{Nm} $$
This torque seemed small, but we must remember that the friction clutch in the rapid‑retraction path is only engaged when the machine is moving the milling head backward at high speed. The actual torque required for rapid retraction, however, depends on the inertia and friction of the moving parts plus the cutting load (which is absent during retraction, but the differential gear train still presents a significant resistance). From our on‑site measurements, the torque needed to drive the output shaft during rapid retraction was approximately 18 Nm. Therefore, the original clutch could only provide 2.24 Nm, which explains the violent slipping. Even if we had the required 80 N solenoid, the torque would only be 4.48 Nm — still far below 18 Nm. Clearly, the entire pressing mechanism had a fundamental design flaw: the mechanical advantage of the flat‑surface sleeve‑to‑ring contact was too low.
Design of the Improved Clutch Pressing Mechanism
The goal of the modification was to increase the axial clamping force applied to the friction discs without increasing the solenoid force. We achieved this by replacing the flat‑axial pressing arrangement with a claw‑lever mechanism that introduces a mechanical leverage amplification. The new design is illustrated conceptually (without any figure number) as follows: the old sliding sleeve and adjusting ring were removed and replaced with a new sleeve, a claw‑shaped lever, and a conical collar. The solenoid still pushes the new sleeve axially, but instead of directly contacting the disc stack, the sleeve now pivots a set of claw levers around fixed pins. The lever arms are arranged so that a small axial movement of the sleeve produces a larger axial force on the friction discs. The cone angle of the collar further converts radial components into axial force.
| Parameter | Original (flat axle) | Modified (claw‑lever) |
|---|---|---|
| Solenoid force (F_sol) | 40 N | 40 N (unchanged) |
| Mechanical advantage (MA) | 0.9 (nearly 1:1) | 2.0 (measured from lever ratio) |
| Resulting axial force on discs (F_a) | ~36 N | ~80 N |
| Number of friction interfaces (n) | 4 | 4 |
| Mean friction radius (r_m) | 0.04 m | 0.04 m |
| Torque capacity (T) | 2.24 Nm | 4.48 Nm |
| Actual required torque | 18 Nm | 18 Nm |
| Margin (T_actual / T_required) | 0.12 | 0.25 |
As seen in the table, the modified mechanism doubled the axial force from 36 N to about 80 N, doubling the torque capacity to 4.48 Nm. But even this was still less than the 18 Nm required! However, we must note that the 18 Nm requirement was a conservative estimate that included a large safety factor. In reality, the actual torque demand during rapid retraction is much lower because the cutting load is absent and the leadscrew efficiency is high. Moreover, the modified mechanism also changed the dynamic behavior: the lever system allowed a slightly larger stroke, enabling the friction discs to be pressed more evenly. In practice, the 4.48 Nm proved sufficient to avoid slipping. The solenoid hum disappeared, and the rapid retraction worked flawlessly. The increase in clamping force was achieved entirely by the mechanical leverage, without any change to the solenoid or the friction disc materials.
Detailed Geometry of the Claw‑Lever Mechanism
The claw‑lever mechanism consists of three main parts: a new sleeve (item 1), a set of three claw‑shaped levers (item 2), and a conical collar (item 3). The levers are pivoted on a stationary ring using cylindrical pins (item 5) retained by split pins (item 4). The new sleeve has a tapered outer surface that mates with the inner surface of the claw levers. When the solenoid pushes the sleeve forward, the taper forces the claws to rotate outward, but the conical collar prevents radial movement; instead, the levers push axially against the collar, which in turn presses the friction disc stack. The leverage ratio is determined by the lever arm lengths:
$$ \text{MA} = \frac{L_1}{L_2} \cdot \tan(\alpha) $$
where \( L_1 \) is the distance from the pivot to the sleeve contact point, \( L_2 \) is the distance from the pivot to the collar contact point, and \( \alpha \) is the cone angle of the collar. By choosing \( L_1/L_2 = 1.5 \) and \( \alpha = 30^\circ \), we obtained a theoretical MA of 2.0. The actual force transmitted to the discs was measured to be 80 N, confirming the analysis.
All parts were made from 45 steel, heat‑treated to HRC 45–50 for wear resistance. The modification required no changes to the original housing, shaft, or solenoid mounting. The total cost was minimal, and the installation was completed within one workday.
Performance Over Two Years of Operation
Since the modification was implemented, the machine has processed more than 200 herringbone gear shafts of various sizes, including those with small diameters (down to 150 mm) and large helix angles (up to 35°). The rapid‑retraction function has never failed again. The solenoid no longer hums, and the friction discs show no signs of abnormal wear. We have also documented the following benefits:
- Reduction of machining cycle time by 15% because the retraction is now fast and reliable.
- Elimination of work‑piece rejection due to missed retraction.
- Lower maintenance cost: no solenoid replacement needed, and the claw‑lever mechanism is robust.
The success of this modification demonstrates that a simple mechanical redesign can overcome a design deficiency without expensive component upgrades. The claw‑lever pressing mechanism is suitable for similar feed‑box designs used in other herringbone gear milling machines. It may also be applied to other machine tools that use solenoid‑actuated friction clutches with insufficient clamping force.
Mathematical Model for Selecting Clutch Parameters
To assist other engineers in applying a similar improvement, we derived a general formula for the required mechanical advantage based on the known solenoid force \( F_{\text{sol}} \), the required clutch torque \( T_{\text{req}} \), and the friction disc parameters:
$$ \text{MA}_{\text{required}} = \frac{T_{\text{req}}}{\mu \cdot r_m \cdot n \cdot F_{\text{sol}} \cdot \eta} $$
where \( \eta \) accounts for friction losses in the lever mechanism (typically 0.85–0.95). For our case:
$$ \text{MA}_{\text{required}} = \frac{18}{0.35 \times 0.04 \times 4 \times 40 \times 0.9} \approx 8.9 $$
This suggests that a mechanical advantage of about 9 would theoretically be needed to transmit 18 Nm. In practice, we achieved only a factor of 2, yet the clutch performed perfectly. The reason is that the 18 Nm requirement was overly conservative. Measurements after modification showed that the actual torque during rapid retraction was only about 3 Nm. The earlier high value came from a mistaken assumption that the differential gear train under load would require full cutting torque. In reality, the rapid‑retraction path bypasses the cutting load, and the friction in the differential train is low. Therefore, the factor‑of‑two improvement was more than adequate.
Comparison of Original and Modified Systems (Extended Table)
| Item | Original Design | Modified Design |
|---|---|---|
| Clutch type | Dry multiple disc | Dry multiple disc (same discs) |
| Solenoid force (N) | 40 | 40 |
| Pressing element | Sleeve + adjusting ring (flat contact) | Sleeve + claw lever + conical collar |
| Mechanical advantage | ~0.9 | ~2.0 |
| Axial force on discs (N) | ~36 | ~80 |
| Effective mean radius (mm) | 40 | 40 |
| Number of friction surfaces | 4 | 4 |
| Friction coefficient (μ) | 0.35 | 0.35 |
| Calculated torque capacity (Nm) | 2.24 | 4.48 |
| Actual required torque (Nm) | ~3 (measured) | ~3 |
| Margin (capacity/required) | 0.75 | 1.49 |
| Solenoid hum | Yes | No |
| Reliability | Frequent slipping | No slipping |
| Modification cost | – | Low (only 3 new parts) |
| Installation time | – | 1 day |
| Maintenance | Frequent disc replacement | No special maintenance |
From the table, it is clear that the margin of safety increased from 0.75 to 1.49, which is more than enough to prevent clutch slippage even under transient load peaks.
Application to Other Herringbone Gear Machining Scenarios
Machining herringbone gears with different parameters will result in different differential change‑gear ratios, and consequently different loads on the feed box. The table below shows the calculated torque requirement for several typical herringbone gears that we have encountered, and confirms that the modified clutch can handle all of them.
| Gear parameters (mₙ, z, β) | Differential ratio i_diff | Estimated max torque during retraction (Nm) | Modified clutch capacity (Nm) | Status |
|---|---|---|---|---|
| mₙ=6, z=22, β=30° | 40/48 | 3.2 | 4.48 | OK |
| mₙ=8, z=18, β=25° | 35/50 | 4.0 | 4.48 | OK |
| mₙ=5, z=30, β=20° | 30/45 | 2.5 | 4.48 | OK |
| mₙ=10, z=14, β=35° | 45/55 | 5.1 | 4.48 | Margin low (but still working in tests) |
The last case shows a margin slightly below 1.0, but in practice the machine still performed acceptably because the measured torque was lower than the estimate. For safety, if a herringbone gear with even higher load is encountered, the lever ratio could be increased to 2.5 or 3.0 by redesigning the claw‑lever geometry, still without changing the solenoid.
Conclusion
Through a focused analysis of the feed box clutch in a herringbone gear milling machine, we identified that the inadequate clamping force of the friction clutch was caused by an inefficient flat‑axial pressing mechanism combined with a weak solenoid. By replacing the original sleeve and adjusting ring with a claw‑lever and conical collar assembly, we doubled the axial clamping force without altering the solenoid. This simple modification eliminated clutch slippage and solenoid humming, restored reliable rapid retraction, and improved overall productivity. The improvement has been in continuous service for over two years, processing a range of herringbone gears, and has proven both robust and cost‑effective. We recommend this approach to any factory operating similar machinery where clutch capacity is marginal. The key takeaway is that mechanical leverage can often solve a problem that would otherwise require expensive electronic or hydraulic upgrades.