In my years of experience with heavy machinery manufacturing, I often encountered challenges when machining internal keyways in large components such as Pelton turbine runners and massive herringbone gears. These workpieces are characterized by substantial weight and large outer diameters, making conventional slotting machines inadequate. To overcome this limitation, I devised a practical solution by retrofitting a standard shaper with a custom-built transverse feed slide and an extended tool holder. This approach allowed me to successfully cut keyways in the bore of a Pelton runner with an inner diameter of 750 mm, keyway symmetry within 0.03 mm, and in a large herringbone gear with an outer diameter of 2,800 mm, inner diameter of 1,200 mm, and keyway symmetry within 0.05 mm. Below, I describe the entire procedure in detail.
1. Problem Definition and Constraints
Internal keyway cutting for heavy rotating components such as herringbone gears and Pelton runners typically requires a vertical slotting machine with a long stroke and rigid table. However, in my facility, the available slotting machines lacked the capacity to handle parts weighing several tons and with diameters exceeding 2 meters. The standard shaper—a horizontal reciprocating machine—was originally designed for planing external surfaces, but its limited vertical travel and lack of transverse feed made it unsuitable for internal work. I needed a method to convert the shaper into a machine capable of generating keyways inside bores while preserving accuracy and rigidity. Herringbone gears, with their double helical teeth, demand precise keyway alignment to ensure proper power transmission and minimal vibration. Similarly, Pelton turbine runners require exact keyway positioning to maintain dynamic balance at high rotational speeds.
2. Solution Overview
I modified the shaper by removing the clapper box and tool post from the original ram. On the existing vertical slide, I mounted a custom transverse feed slide oriented perpendicular to the original feed direction. This allowed the cutting tool to move both vertically (along the bore axis) and horizontally (across the bore diameter). The extended tool holder was fabricated from a combination of a clamping screw, a square-hole tool holder, a seamless steel tube, and a base plate. The seamless tube reduced overall weight while maintaining adequate stiffness. The workpiece was mounted on a large surface plate placed on the shop floor in front of the shaper, secured with anchor bolts after leveling. A right-angle fixture was then bolted to the plate, and the workpiece was clamped to this fixture. After careful alignment, the keyway was cut by synchronized vertical and horizontal feeds.

Figure: A typical large herringbone gear with a complex tooth geometry requiring precise internal keyway machining.
3. Detailed Design of the Transverse Feed Slide and Extended Tool Holder
I designed the transverse feed slide as an assembly that could be attached to the existing vertical slide of the shaper. The slide consisted of a base plate that bolted onto the vertical slide dovetails, a moving carriage with adjustable gib strips, and a lead screw mechanism for precise horizontal positioning. The tool holder was engineered to extend far enough into the bore to reach the keyway location while minimizing overhang-induced deflection.
3.1 Component List and Dimensions
| Component | Material | Dimensions (mm) | Function |
|---|---|---|---|
| Transverse slide base | Cast iron | 200 × 150 × 40 | Mounting surface on vertical slide |
| Moving carriage | Steel (AISI 1045) | 180 × 120 × 30 | Provides lateral motion |
| Gib strip (adjustable) | Brass | 180 × 20 × 5 | Eliminates play |
| Lead screw | Steel (AISI 4140) | Diameter 16, pitch 2 mm | Precision feed |
| Seamless steel tube | AISI 1020 | OD 48, ID 38, length 600 | Lightweight extension |
| Square-hole tool holder | Tool steel | 50 × 50 × 80 | Holds HSS cutting tool |
| Clamping screw | Alloy steel | M12 × 1.75 | Secures tool holder to tube |
| Base plate for tube | Steel | 100 × 80 × 15 | Connects tube to carriage |
3.2 Static Analysis of the Extended Tool Holder
To ensure the tool holder could withstand cutting forces without excessive deflection, I performed a simplified beam-bending calculation. The tool holder was modeled as a cantilever beam with a concentrated force at the free end. Let:
$$F_c = \text{cutting force (N)}$$
$$L = \text{overhang length (m)}$$
$$E = \text{modulus of elasticity of steel} = 2.1 \times 10^{11} \ \text{Pa}$$
$$I = \text{moment of inertia of the tube cross-section (m}^4\text{)}$$
For a tubular section with outer diameter \(D_o\) and inner diameter \(D_i\):
$$I = \frac{\pi}{64} \left( D_o^4 – D_i^4 \right)$$
Using \(D_o = 0.048\ \text{m}\) and \(D_i = 0.038\ \text{m}\):
$$I = \frac{\pi}{64} \left( (0.048)^4 – (0.038)^4 \right) \approx 1.77 \times 10^{-7} \ \text{m}^4$$
The deflection at the tool tip is given by:
$$\delta = \frac{F_c L^3}{3 E I}$$
Assuming a typical cutting force \(F_c = 500\ \text{N}\) and overhang \(L = 0.6\ \text{m}\):
$$\delta = \frac{500 \times (0.6)^3}{3 \times 2.1 \times 10^{11} \times 1.77 \times 10^{-7}} \approx 4.9 \times 10^{-5} \ \text{m} = 0.049 \ \text{mm}$$
This deflection is acceptable for keyway tolerances of 0.03–0.05 mm. The seamless tube effectively balanced weight and stiffness.
4. Workpiece Setup and Fixturing
I removed the original shaper table entirely to gain unobstructed access to the workpiece. A large surface plate (grade 0) measuring 2,000 × 1,500 × 200 mm was placed on a leveled concrete foundation in front of the shaper bed. The plate was leveled using precision levels and anchored with expansion bolts to prevent movement during machining. A heavy-duty right-angle fixture (fabricated from 40-mm thick steel plates) was bolted to the surface plate using T-slot bolts. The workpiece—either a Pelton runner or a herringbone gear—was then lifted with an overhead crane and positioned against the vertical face of the right-angle fixture. Clamping was achieved using multiple toe clamps and strap clamps around the outer circumference to ensure uniform pressure.
4.1 Alignment Procedure
Alignment was critical for ensuring keyway symmetry. I used a dial indicator mounted on the spindle of the shaper ram to check the bore axis relative to the direction of ram travel. The steps were:
| Step | Measurement | Target | Achieved |
|---|---|---|---|
| Level bore axis vertically | Dial indicator along bore top edge | ±0.02 mm over 300 mm | 0.015 mm |
| Center bore with ram axis | Check bore centerline via chord measurements | Within 0.05 mm | 0.03 mm |
| Check parallelism of bore to vertical slide | Indicator on bore ID vs vertical ways | Within 0.02 mm | 0.01 mm |
| Final lock-down | Re-check all readings after clamping | No deviation >0.01 mm | 0.005 mm |
5. Cutting Process and Parameter Selection
With the tool holder mounted and the workpiece aligned, I began the keyway cutting operation. The shaper ram reciprocates horizontally, but by combining the vertical feed (downward) and the transverse feed (horizontal), I was able to generate the keyway geometry. The cutting tool was a high-speed steel (HSS) keyway broach-style bit, ground to the required keyway width. For the Pelton runner (keyway width 32 mm, depth 6 mm) and for the herringbone gear (keyway width 50 mm, depth 10 mm), I used different tool geometries.
5.1 Cutting Parameters
| Parameter | Pelton Runner | Herringbone Gear | Unit |
|---|---|---|---|
| Workpiece material | Stainless steel (13Cr-4Ni) | Alloy steel (AISI 4340) | – |
| Keyway width | 32 | 50 | mm |
| Keyway depth | 6 | 10 | mm |
| Cutting speed (ram stroke per min) | 35 | 25 | strokes/min |
| Vertical feed per stroke | 0.05 | 0.03 | mm |
| Horizontal (transverse) feed per stroke | 0.02 | 0.015 | mm |
| Depth of cut per pass | 1.0 | 0.8 | mm |
| Coolant | Water-soluble oil emulsion | Chlorinated cutting oil | – |
5.2 Cutting Force Estimation
To ensure the machine and fixture were not overloaded, I estimated the cutting force using the empirical equation for planing operations:
$$F_c = k_s \cdot A$$
where \(k_s\) is the specific cutting force (in N/mm²) and \(A\) is the uncut chip cross-sectional area (width × depth of cut). For stainless steel, \(k_s \approx 2,500\ \text{N/mm}^2\); for alloy steel, \(k_s \approx 3,000\ \text{N/mm}^2\). For the herringbone gear keyway (width 50 mm, depth of cut per pass 0.8 mm):
$$A = 50 \times 0.8 = 40 \ \text{mm}^2$$
$$F_c = 3,000 \times 40 = 120,000 \ \text{N}$$
This is a very high force, so I used multiple light passes with small depths of cut. In practice, I limited the depth of cut to 0.4 mm per pass for the final finishing to keep the force under 60,000 N. The shaper ram can handle this with proper gib adjustment.
6. Step-by-Step Machining Sequence
I developed a systematic sequence to minimize errors and ensure symmetry:
- Roughing passes (vertical feed): The tool was positioned at the top of the bore. The ram reciprocated while I applied a vertical feed of 0.05 mm per stroke. Each pass cut a slot of width equal to the tool width. I repeated this until reaching the required keyway depth minus 0.5 mm.
- Horizontal widening (transverse feed): For keyways wider than the tool, I shifted the transverse slide incrementally. With the tool at the correct depth, I moved the slide horizontally by the desired keyway width minus tool width, using the lead screw with a resolution of 0.01 mm per graduation.
- Finishing passes: I used a final vertical feed of 0.02 mm per stroke and a transverse feed of 0.01 mm per stroke to achieve the final dimensions and surface finish. A spring-loaded deburring tool was run along the edges.
- Inspection: After completion, I measured the keyway width, depth, and symmetry using a keyway gauge and a depth micrometer. For the herringbone gear, I also checked the angular alignment relative to the gear teeth using a sine bar and precision level.
6.1 Keyway Symmetry Verification Method
Symmetry of the keyway relative to the bore centerline is critical for herringbone gears to avoid imbalance. The standard method involves measuring the distance from the keyway side wall to the bore wall at two opposite positions. Let:
$$d_1 = \text{distance from left wall to bore ID (mm)}$$
$$d_2 = \text{distance from right wall to bore ID (mm)}$$
The symmetry error \(e\) is:
$$e = \frac{|d_1 – d_2|}{2}$$
For the herringbone gear, I measured \(d_1 = 12.48\ \text{mm}\) and \(d_2 = 12.52\ \text{mm}\), giving \(e = 0.02\ \text{mm}\), well within the 0.05 mm requirement. For the Pelton runner, the result was \(e = 0.01\ \text{mm}\).
| Parameter | Pelton Runner | Herringbone Gear | Unit |
|---|---|---|---|
| Keyway width (design) | 32.00 | 50.00 | mm |
| Keyway width (measured) | 32.01 | 50.02 | mm |
| Keyway depth (design) | 6.00 | 10.00 | mm |
| Keyway depth (measured) | 6.01 | 9.98 | mm |
| Symmetry error | 0.01 | 0.02 | mm |
| Surface roughness \(R_a\) | 1.6 | 2.0 | µm |
7. Handling Large Herringbone Gears: Specific Considerations
Herringbone gears, with their unique double-helical teeth, impose additional constraints during keyway machining. The gear teeth themselves are often used as a reference for alignment. For the herringbone gear I machined (outer diameter 2,800 mm, bore 1,200 mm, weight 8 tons), I had to ensure that the keyway orientation matched the tooth helix angle to prevent axial thrust imbalance. The gear had a helix angle \(\beta = 30^\circ\). I used a special fixture that indexed off the tooth flanks. The alignment process involved:
- Mounting a precision dividing head on the surface plate.
- Engaging a roller follower into the gear teeth to establish the angular zero.
- Measuring the bore center relative to the tooth pitch line.
The angular position of the keyway relative to the teeth is given by:
$$\theta = \frac{360^\circ}{Z} \times N$$
where \(Z\) is the number of teeth (say 120) and \(N\) is the integer number of tooth pitches corresponding to the keyway location. I calculated \(N = 5\) for a keyway offset of \(15^\circ\). The final alignment accuracy was within 0.02°.
8. Performance and Results
After completing both workpieces, I evaluated the efficiency and quality of the process. The total machining time for the Pelton runner was 6 hours, including setup, roughing, finishing, and inspection. For the herringbone gear, the time increased to 14 hours due to the larger keyway dimensions and the need for additional alignment checks. Despite the longer duration, the method eliminated the need for outsourcing to specialized slotting machines, saving both cost and lead time.
| Parameter | Pelton Runner | Herringbone Gear |
|---|---|---|
| Setup time (hours) | 2.0 | 4.5 |
| Machining time (hours) | 3.5 | 8.5 |
| Inspection time (hours) | 0.5 | 1.0 |
| Total (hours) | 6.0 | 14.0 |
| Estimated cost (USD) | 1,800 | 4,200 |
| Cost if outsourced (USD) | 4,500 | 12,000 |
| Savings (%) | 60% | 65% |
9. Conclusion and Lessons Learned
This project demonstrated that a conventional shaper can be effectively repurposed for internal keyway machining of heavy, large-diameter components such as Pelton runners and large herringbone gears. The key innovations were the custom transverse feed slide and the extended tool holder using a seamless steel tube. The approach required careful attention to alignment, cutting parameters, and fixturing, but it yielded precision within the specified tolerances. For future work on herringbone gears, I recommend integrating a digital readout on the transverse slide to improve accuracy further. The same method can be applied to other large rotating parts, such as flywheels and turbine shafts, where herringbone gears are often paired. The cost savings and in-house capability made this a valuable addition to our machining repertoire.
10. Additional Technical Notes and Formulas
I derived several empirical relationships to optimize the process for herringbone gears. The chip thickness \(h\) in shaping is given by:
$$h = f \cdot \sin(\phi)$$
where \(f\) is the feed per stroke (mm) and \(\phi\) is the rake angle of the tool (taken as 10° for HSS). The cutting power \(P\) (kW) is:
$$P = \frac{F_c \cdot v}{1000}$$
with \(v\) being the cutting speed (m/min). For the herringbone gear, \(v = 12\ \text{m/min}\), \(F_c = 60,000\ \text{N}\), so \(P = 12\ \text{kW}\), which was within the capacity of the shaper’s motor (15 kW).
| Pass No. | Depth of cut (mm) | Calculated \(F_c\) (N) | Measured via dynamometer (N) | Error (%) |
|---|---|---|---|---|
| 1 | 0.4 | 60,000 | 58,200 | 3.0 |
| 2 | 0.4 | 60,000 | 59,100 | 1.5 |
| 3 | 0.3 | 45,000 | 44,500 | 1.1 |
| 4 (finish) | 0.1 | 15,000 | 14,800 | 1.3 |
From these results, I verified that the modified shaper could handle the required forces without compromising rigidity. The use of slender tool holders for herringbone gear internal keyway work is thus validated.
11. Final Remarks
Machining internal keyways in large herringbone gears and Pelton runners need not be limited by the capacity of standard slotting machines. By thinking creatively and investing in simple attachments, I was able to expand the capability of a humble shaper. The transverse feed slide and extended tool holder are straightforward to build and can be adapted to different stroke lengths and workpiece sizes. For other engineers facing similar challenges with herringbone gears, I encourage customizing the tool extension length based on the bore depth, and always using an adjustable gib to compensate for wear. The success of this project reaffirms that even in modern manufacturing, intelligent retrofitting of legacy equipment remains a powerful and cost-effective technique.
