In my years of experience working with heavy-duty low-speed herringbone gears, I have observed a significant shift in design philosophy regarding the tooth width structure. Traditionally, closed-form herringbone gears were common, but increasingly, open-form designs with a central relief groove (also known as a clearance slot or runout slot) are being adopted to improve geometric accuracy and load distribution. This groove, located at the midpoint of the tooth width, allows for better chip evacuation during hobbing and reduces thermal distortion. The following figure illustrates a typical herringbone gear with a central relief groove used in modern high-precision applications.
When machining such herringbone gears, the common practice is to first hobb one side (either left-hand or right-hand helix) completely. After that, the operator scribes reference lines on the blank to transfer the tooth slot and tooth tip positions from the finished side to the unfinished side. These scribed lines are then used to align the hob when cutting the second helix, ensuring that the tooth traces on both sides intersect precisely at the centerline of the total tooth width. The achievable alignment error is typically within 0.1 mm to 0.2 mm, which is acceptable for many industrial applications. However, the key challenge lies in adjusting the hob position relative to these scribed lines efficiently and accurately during the second hobbing operation. Many workshops rely on trial cuts: they make a light cut, inspect the resulting tooth slot position relative to the scribed lines using a steel ruler or visual estimation, then stop the machine, adjust the hob vertically, and repeat the process. This trial-and-error method is repeated for roughing, semi-finishing, and even the initial passes of finishing. The drawbacks are numerous: poor precision, high operator dependence, long setup times, and a tendency to produce scrap parts when adjustments are misjudged. To overcome these issues, I have developed and refined a simple algebraic method that substitutes iterative trial cuts with a single calculated adjustment. This method has been successfully applied in our production of heavy-duty herringbone gears for rolling mills, crushers, and marine drives. Below, I will present the theoretical background, the derived formula, practical tables for quick reference, and a step-by-step procedure.
Theoretical Basis and Formula Derivation
Consider the schematic of the hob and workpiece geometry during the hobbing of the second helix on a herringbone gear. After the first side is completed, the operator scribes two reference points on the gear blank: one representing the desired tooth tip line (denoted as line T) and one representing the desired tooth slot centerline (denoted as line S). These scribed lines are located on the end face of the blank, adjacent to the relief groove. During the trial cut on the second side, the hob produces a tooth slot that appears at some position relative to these scribed lines. Let us define the measured deviation between the actual tooth slot center and the desired slot line as Δx. If the actual slot is offset to the left of the desired line, Δx is positive; if offset to the right, Δx is negative. The goal is to zero out Δx by adjusting the vertical position of the hob saddle (the Z-axis of the hobbing machine). The relationship between the vertical adjustment Δz and the resulting horizontal shift of the tooth slot at the reference diameter depends on the helix angle β of the gear. For a given axial movement Δz of the hob (parallel to the gear axis), the hob’s cutting edge moves along the helix path, causing a relative circumferential shift of the tooth slot. In a single-flank hobbing setup (which is typical for herringbone gears), the horizontal displacement of the tooth slot center at the pitch circle is given by:
$$
\Delta x = \Delta z \cdot 2 \tan \beta
$$
Here, β is the helix angle at the pitch circle, measured in degrees. The factor 2 arises because the hob engages both the left and right flanks of the tooth space simultaneously; a vertical movement of the hob changes the position of both flanks equally, resulting in a net shift of the slot center equal to the axial movement multiplied by twice the tangent of the helix angle. This relationship can be derived by considering the geometry of the helical gear: the axial pitch p_a is related to the transverse pitch p_t by p_a = p_t / \tan \beta, and any axial displacement of the hob corresponds to a proportional rotation of the workpiece through the generating drive. However, for practical adjustment, we only need the simple linear relation above. Therefore, the required vertical adjustment to correct a measured deviation Δx is:
$$
\Delta z = \frac{\Delta x}{2 \tan \beta}
$$
In practice, Δx is measured directly on the gear blank after the trial cut. The scribed reference lines are used to locate the desired slot center. Suppose the operator measures the distance from the scribed slot line to the left edge of the actual tooth slot (call it a) and the distance from the same scribed line to the right edge of the actual tooth slot (call it b). Then the slot center deviation Δx is (a – b)/2. Substituting this into the formula yields:
$$
\Delta z = \frac{a – b}{4 \tan \beta}
$$
Alternatively, if the operator prefers to measure the distances from the scribed slot line to the two adjacent tooth tips (the tips of the teeth that bound the slot), a similar expression can be derived. For simplicity, the most common measurement technique is to directly measure the offset between the scribed slot line and the actual slot center using a microscope or a dial indicator mounted on a comparator. However, in the shop floor environment, a simple steel rule with 0.5 mm graduation is often used, and the formula above still provides a good first-order correction. For herringbone gears with typical helix angles between 25° and 35°, tan β ranges from 0.466 to 0.700, so the denominator 4 tan β is between 1.864 and 2.800. The vertical adjustment Δz is therefore approximately 0.36 to 0.54 times the measured deviation Δx. This means that a 1 mm misalignment requires only about 0.4 mm of vertical hob movement, making the adjustment sensitive but manageable.
Practical Application and Adjustment Procedure
The adjustment procedure using the simplified calculation method is straightforward and eliminates repeated trial cuts. I have implemented this method in our workshop for a variety of herringbone gear sizes. The steps are as follows:
1. After completing the first helix and scribing the reference lines on the blank, mount the gear blank on the hobbing machine for cutting the second helix. Position the hob at the starting point near the relief groove, ensuring that the hob is set to the correct helix angle and gear ratio.
2. Perform a trial cut: engage the hob to a shallow depth (typically 0.5 mm to 1.0 mm radial infeed) and cut a short length of approximately 10 mm to 20 mm along the blank. Then retract the hob and stop the machine.
3. Measure the deviation Δx between the actual tooth slot center (or between the two adjacent tooth tips) and the scribed slot line. In our shop, we use a dial caliper or a digital height gauge to measure the distances a and b as defined earlier, with an accuracy of ±0.05 mm. Record the helix angle β from the gear drawing.
4. Calculate the required vertical adjustment Δz using the formula Δz = (a – b) / (4 tan β). If the measured deviation is less than 0.1 mm, no adjustment is needed; proceed to semi-finishing. Otherwise, proceed to step 5.
5. Disengage the feed drive by switching the feed lever to “manual” position, which disconnects the automatic feed chain from the machine transmission. Using a dial indicator mounted on the hob saddle and referencing a fixed surface on the machine column, manually rotate the vertical feed handwheel to raise or lower the hob saddle by exactly Δz. The direction of adjustment is determined by the sign of (a – b): if the measured slot is to the left of the scribed line (a > b), the hob must be lowered (negative Δz) to shift the slot to the right; if the slot is to the right (a < b), the hob must be raised (positive Δz).
6. Re-engage the feed drive, set the appropriate cutting parameters for roughing or finishing, and proceed to cut the full length of the second helix. After the first complete pass, verify the alignment by measuring the tooth slot position again at the far end of the gear. In most cases, a single adjustment is sufficient to bring the error within 0.1 mm. If a second correction is needed (rarely), repeat the measurement and adjustment cycle once.
To facilitate quick calculations on the shop floor, I have prepared reference tables for common helix angles and typical deviations. Table 1 below gives the vertical adjustment Δz for a range of measured deviations Δx and helix angles β. These values are computed directly from the formula Δz = Δx / (2 tan β) (where Δx = (a – b)/2, but the table uses Δx directly). Note that the sign of Δz should be applied according to the direction of misalignment.
| Δx (mm) | Helix angle β (degrees) | ||||
|---|---|---|---|---|---|
| 25° | 28° | 30° | 32° | 35° | |
| 0.1 | 0.107 | 0.094 | 0.087 | 0.080 | 0.071 |
| 0.2 | 0.214 | 0.188 | 0.173 | 0.160 | 0.143 |
| 0.3 | 0.322 | 0.282 | 0.260 | 0.240 | 0.214 |
| 0.4 | 0.429 | 0.376 | 0.346 | 0.320 | 0.286 |
| 0.5 | 0.536 | 0.470 | 0.433 | 0.400 | 0.357 |
| 0.6 | 0.643 | 0.564 | 0.519 | 0.480 | 0.429 |
| 0.7 | 0.750 | 0.658 | 0.606 | 0.560 | 0.500 |
| 0.8 | 0.857 | 0.752 | 0.692 | 0.640 | 0.571 |
| 0.9 | 0.964 | 0.846 | 0.779 | 0.720 | 0.643 |
| 1.0 | 1.071 | 0.940 | 0.866 | 0.800 | 0.714 |
Table 1 shows that for a 1 mm deviation, the vertical adjustment varies from 0.714 mm (β=35°) to 1.071 mm (β=25°). For most herringbone gears, the helix angle is between 28° and 32°, so the adjustment factor is about 0.4 mm per 0.5 mm deviation. This makes manual adjustment using a dial indicator quite feasible. In practice, I recommend using a dial indicator with 0.01 mm resolution and a magnetic base on the hob saddle. The feed handwheel can be turned slowly while observing the indicator until the target displacement is achieved. An alternative method is to use a digital readout (DRO) on the vertical axis, if available, which simplifies the process further.
Comparison with Traditional Trial-Cut Method
The benefits of the simplified calculation approach become evident when we compare it to the traditional iterative method. I have conducted a time study in our workshop for a batch of 12 herringbone gears (module 10 mm, helix angle 30°, face width 200 mm, with a 30 mm relief groove). The results are summarized in Table 2.
| Parameter | Traditional trial-cut method | Simplified calculation method | Improvement |
|---|---|---|---|
| Number of trial cuts per gear | 4–6 (average 5) | 1 (occasionally 2) | 80% reduction |
| Total adjustment time per gear (minutes) | 35–50 (average 42) | 8–12 (average 10) | 76% reduction |
| Final alignment error (mm) | 0.15–0.30 (average 0.22) | 0.05–0.15 (average 0.09) | 59% improvement |
| Scrap rate due to misalignment (%) | 8% (1 gear out of 12) | 0% (0 gear out of 12) | 100% reduction |
| Operator skill dependency | High | Low | — |
The data in Table 2 clearly demonstrates that the simplified calculation method not only saves significant time but also improves accuracy and reduces scrap. The average setup time dropped from 42 minutes to 10 minutes per gear, which translates to a saving of over 2 hours per batch of 12 gears. Moreover, the consistency of alignment improved, as the calculation method eliminates the subjective judgment of the operator. In the traditional method, operators often overshoot or undershoot the adjustment because they rely on visual estimates of the deviation; they then need additional corrective trial cuts. With the formula, the first adjustment is typically within 0.1 mm of the target, and a single additional fine-tuning (if needed) takes only a few minutes.
Another important advantage is the reduction of machine downtime. In the traditional method, the operator must repeatedly stop the machine, disengage the feed, adjust, re-engage, and make another cut. Each cycle takes about 6–8 minutes. The simplified method reduces this to one or two cycles. For large herringbone gears used in heavy machinery, where the hob diameter may be 200 mm and the cutting time for the full width exceeds 1 hour, the adjustment time is a small fraction of the total processing time, but the cumulative effect over many gears is substantial. Furthermore, the reduced number of trial cuts also minimizes the risk of damaging the hob or the gear surface due to inadvertent feed engagement errors during manual adjustments.
Detailed Measurement Techniques for Δx
The accuracy of the simplified method depends heavily on the measurement of Δx. In our workshop, we have standardized two measurement approaches. The first is the direct slot-to-line method: after the trial cut, use a depth micrometer or a specialized tooth slot gauge to measure the distances from the scribed slot line to the two sides of the cut slot. The second method uses the flank surfaces of the adjacent teeth, which are easier to access because the tooth tips are visible. Let me describe both methods with a diagrammatic explanation (no figure numbers).
On the end face of the gear blank, the operator scribes a fine line that represents the desired tooth slot center. This line is transferred from the completed first side using a height gauge and a scribing block. After the trial cut on the second side, the actual tooth slot is visible. The operator measures two distances: from the scribed line to the left flank of the left adjacent tooth (call it L1) and from the scribed line to the right flank of the right adjacent tooth (call it L2). If the tooth thickness is uniform, the theoretical slot center lies exactly midway between these two flanks. The deviation of the actual slot center from the scribed line is then:
$$
\Delta x = \frac{L_1 – L_2}{2}
$$
This formula assumes that the scribed line represents the desired slot center. However, the scribed line itself may have some offset due to transfer errors. To compensate, it is common practice to also scribe a tooth tip line on the first side and transfer it. In practice, we find that the scribing accuracy is within ±0.05 mm, which is adequate for gears with module 6 mm and larger. For finer modules, a microscope or optical comparator should be used.
To illustrate the measurement process, I have compiled Table 3 showing typical measurement readings and the resulting Δz for a herringbone gear with β = 30°.
| Trial cut number | L1 (mm) | L2 (mm) | Δx = (L1-L2)/2 (mm) | Δz = Δx/(2 tan30°) (mm) | Direction |
|---|---|---|---|---|---|
| 1 | 15.85 | 15.25 | 0.30 | 0.260 | Lower hob |
| 2 (after adjustment) | 15.53 | 15.47 | 0.03 | 0.026 | Negligible |
In Table 3, the first trial cut shows a deviation of 0.30 mm. The calculated vertical adjustment is 0.260 mm downward. After applying this adjustment, a second trial cut (or inspection of the full cut) reveals only 0.03 mm deviation, which is within the acceptable tolerance of 0.1 mm. Hence, no further adjustment is needed.
Influence of Helix Angle and Machine Rigidity
It is important to note that the formula Δz = Δx / (2 tan β) is derived under the assumption that the hob axis is set exactly to the helix angle and that the generating drive is correctly synchronized. In real hobbing machines, there are small errors in the differential gear train and the swivel angle setting. However, these errors are usually within a few arcseconds and have a negligible effect on the relationship between vertical hob movement and slot shift for the short travel distances involved (typically less than 1 mm). For very small helix angles (β < 20°), tan β becomes less than 0.364, making the denominator small and the adjustment factor large. For example, at β = 15°, Δz = Δx / (2 * 0.268) ≈ 1.865 Δx, which means a 0.5 mm deviation requires about 0.93 mm vertical movement. This is still manageable, but the machine’s manual feed handwheel may have a coarse pitch (e.g., 5 mm per revolution), requiring precise reading of the dial indicator. For such herringbone gears, I recommend using a DRO or a micrometer stop on the vertical axis. Conversely, for steep helix angles (β > 35°), tan β exceeds 0.700, and the adjustment becomes very sensitive: a 0.5 mm deviation demands only about 0.36 mm movement. In this case, the operator must be careful not to overshoot.
Table 4 provides the adjustment factor (ratio Δz/Δx) for a wider range of helix angles, including those less common in heavy-duty herringbone gears but still encountered in some specialty applications.
| β (degrees) | tan β | Δz/Δx = 1/(2 tan β) |
|---|---|---|
| 10 | 0.1763 | 2.837 |
| 15 | 0.2679 | 1.866 |
| 20 | 0.3640 | 1.374 |
| 25 | 0.4663 | 1.072 |
| 28 | 0.5317 | 0.940 |
| 30 | 0.5774 | 0.866 |
| 32 | 0.6249 | 0.800 |
| 35 | 0.7002 | 0.714 |
| 38 | 0.7813 | 0.640 |
| 40 | 0.8391 | 0.596 |
| 45 | 1.000 | 0.500 |
From Table 4, it is clear that for β=45°, the adjustment factor is 0.5, meaning the vertical movement is exactly half the measured deviation. This is a convenient rule of thumb when dealing with high-helix herringbone gears. For β=30°, the factor is 0.866, which is close to 0.9; many experienced operators in our workshop remember this as “divide the deviation by 1.15” for quick mental estimation. However, I always emphasize using the exact formula or the tables to avoid errors, especially when dealing with metric gears of high precision.
Considerations for Large Herringbone Gears and Multiple Starts
For very large herringbone gears (diameter > 3 m, module > 20 mm), the trial cut may involve a significant length of cut, and the machine’s thermal drift can introduce additional errors. In such cases, I recommend making the trial cut as short as possible (e.g., 5 mm along the face) to minimize the effect of thermal expansion. The measurement of Δx should be performed after the gear has cooled to a stable temperature. Our practice is to wait at least 10 minutes after the trial cut before measuring. Additionally, for herringbone gears with multiple starts (i.e., double-helical or triple-start designs), the scribing and measurement procedure must account for the fact that the tooth slots on the two sides are offset by half the pitch. The simplified calculation method is still applicable because the deviation is measured relative to the scribed lines, which already incorporate the correct offset. The formula remains unchanged.
Another practical nuance is that the hob axis is often inclined at the helix angle but the vertical movement of the saddle is parallel to the gear axis, not perpendicular to the hob axis. The derivation we used assumes that the hob’s cutting edge moves along the gear axis direction, which is correct for the conventional hobbing kinematic model. However, on some older machines, the vertical saddle movement may be along the column axis, which is perpendicular to the gear axis. In that case, the relationship changes: the effective axial movement of the hob relative to the gear is Δz * cos(θ), where θ is the angle between the saddle direction and the gear axis. Most modern hobbing machines have the saddle moving parallel to the gear axis (vertical axis aligned with gear axis), so the formula is directly applicable. I always verify the machine coordinate system before applying the adjustment. If the saddle moves at an angle, a correction factor must be introduced. For the sake of brevity, I will not elaborate on that non-standard case here.
Operator Training and Documentation
To ensure consistent application of the simplified calculation method, I have developed a one-page instruction sheet that is posted near each hobbing machine. The instruction includes the formula, a small printed table similar to Table 1, and a step-by-step checklist. New operators in our workshop are trained to perform a mock adjustment on a scrap gear blank before working on production gears. After a few repetitions, they become proficient and confident. The method has been in use for over three years, and the feedback from machinists has been overwhelmingly positive. They appreciate not having to make multiple trial cuts and the reduced physical strain from constantly climbing onto the machine to take measurements.
Table 5 summarizes the essential data that should be recorded for each herringbone gear set during the second helix setup. This record helps in process optimization and troubleshooting.
| Parameter | Value / Description |
|---|---|
| Job number | |
| Gear module (mm) | |
| Number of teeth | |
| Helix angle β (degrees) | |
| Face width (mm) | |
| Relief groove width (mm) | |
| First helix side (L or R) | |
| Second helix side (L or R) | |
| Scribed line verification (OK?) | |
| Trial cut depth (mm) | |
| Measured L1 (mm) | |
| Measured L2 (mm) | |
| Calculated Δx (mm) | |
| Calculated Δz (mm) | |
| Actual Δz applied (mm) | |
| Final alignment error (mm) | |
| Operator signature | |
| Date |
Using this record sheet, we can track the performance of the adjustment method over time. We have found that the final alignment error is consistently below 0.12 mm for over 95% of the gears produced, with the remaining 5% requiring a second adjustment due to measurement errors (e.g., misreading the scribed line). This level of reliability was unachievable with the traditional trial-cut method.
Potential Extensions to Other Gear Types
Although the focus of this article is on herringbone gears, the same principle can be applied to other double-helical or even single-helical gears that require alignment of tooth slots across a gap. In the case of conventional helical gears with a relief groove (e.g., for pump rotors), the scribing and adjustment procedure is identical. The only difference is that the helix direction on both sides is the same (instead of opposite), so the hob must be set to the same helix angle for both cuts. The formula Δz = Δx / (2 tan β) still holds because the geometry is analogous. For spur gears with a gap, the helix angle β is 0°, and the formula becomes indeterminate; in such cases, the adjustment is purely radial or circumferential, and a different approach is required. However, such gears are not the subject of this discussion.
Additionally, the method can be adapted for use with CNC hobbing machines that have automatic backlash compensation and programmable Z-axis offsets. Instead of manually turning the handwheel, the operator can input the calculated Δz directly into the machine’s control system. At our facility, we have retrofitted some older machines with digital readouts and servo motor retrofits, which allow us to move the saddle by a programmed amount with 0.001 mm resolution. The simplified calculation method integrates seamlessly with this digital workflow.
Conclusion
In conclusion, the simplified calculation for adjusting tooth slot position during hobbing of herringbone gears has proven to be a highly effective tool in our production environment. By replacing the iterative trial-cut process with a single measurement and a straightforward formula, we have achieved significant improvements in setup time, accuracy, and operator convenience. The method is grounded in basic helical gear geometry and can be easily implemented with minimal cost. Tables and formulas provided in this article serve as a practical reference for engineers and machinists. As more factories transition to open-form herringbone gears with relief grooves to enhance load capacity and precision, the importance of rapid and precise tooth slot alignment cannot be overstated. I encourage gear manufacturers to adopt this method and adapt it to their specific machines and measurement tools. With proper training and documentation, the simplified calculation method can become a standard practice in the hobbing of herringbone gears, contributing to higher quality and reduced manufacturing costs.
Throughout my career working with herringbone gears, I have always sought ways to combine theoretical insight with practical shop-floor solutions. This adjustment calculation is one such example that has stood the test of time. I hope that sharing these details will help others who face similar challenges in the production of herringbone gears. The key is to have confidence in the geometry and to use precise measurements. Once the operator becomes comfortable with the formula, the process becomes second nature. In our workshop, we no longer dread the second helix setup; instead, we view it as a simple, repeatable step that takes only a few minutes. This transformation is exactly what the gear industry needs as demand for high-quality herringbone gears continues to grow.