Gear Sharpening Technology: How Hard Finishing Works and Where It Is Going

• 1
  Key takeaways
• 2
  What gear sharpening technology actually removes
• 3
  Form grinding, generating grinding, and gear honing compared
• 4
  When hard finishing helps and when it does not
• 5
  Microgeometry, contact pattern, and gear noise
• 6
  Development trends in gear sharpening technology
• 7
  Choosing a flank finishing process

Gear sharpening technology covers the finishing operations that cut or abrade the loaded tooth flank after heat treatment. In the shop, the term gets used loosely for grinding, honing, and profile grinding of case-hardened gears. The common thread is that the tool removes a thin layer of flank material to correct geometry that was distorted by carburizing or nitriding.

The cut is shallow by design. A typical hard-finishing pass removes 0.02 to 0.10 mm from the flank, measured normal to the surface. That range is enough to clean up distortion and leave a defined surface, but it will not rescue a gear that was cut to the wrong module or helix angle.

Hardness decides what the abrasive can do. Case-hardened gears at 58 to 62 HRC respond to cubic boron nitride or ceramic tooling. Below roughly 45 HRC, a skiving or shaving cutter still has an economic edge because it removes material faster and the tool lasts longer.

The flank is only one variable. Runout, bore-to-pitch concentricity, and axial location also set the final mesh pattern. A perfectly sharpened flank on a gear with 0.03 mm of bore runout still runs loud. We check the datum before quoting any finishing pass.

• 1
  Profile correctionRemoves pressure-angle drift caused by heat-treat distortion.
• 2
  Lead correctionStraightens helix error and twist along the face width.
• 3
  Surface definitionSets flank roughness so the oil film can carry load.
• 4
  Not a size fixIt will not correct a wrong module, number of teeth, or center distance.

Form grinding uses a profiled wheel dressed to the gap between two teeth. It cuts the whole flank in one pass and holds profile very tightly, which suits large modules and low-volume work. Setup is slow because the wheel must be dressed for each gear geometry, and the wheel form wears as it cuts.

Generating grinding rolls a threaded or profile wheel against the gear like a rack meshing with the part. The machine produces the involute by motion rather than by wheel shape, so one wheel covers a family of gears. Cycle time is longer per tooth, but changeover between similar parts is short. This is the usual choice for automotive and EV transmission gears.

Gear honing is a low-force abrasive process that runs after grinding. An internal or external honing ring with abrasive grains slides along the flank at crossed axes. It removes only 5 to 20 μm, mainly to reduce roughness and to break up grinding burn marks. Honing does not correct large profile errors, so it cannot replace a grinding pass.

The three processes are not interchangeable. Grinding sets geometry. Honing polishes the surface and nudges microgeometry. Choosing honing when the profile is 30 μm out of tolerance just buries the error under a smoother surface.

• 1
  Form grindingTight profile, large module, low to medium volume.
• 2
  Generating grindingFlexible geometry, high volume, fast changeover.
• 3
  Gear honingMicrogeometry and roughness only, 5–20 μm stock.

Finishing pays off when mesh contact is the limiting factor. High-speed stages, e-transmission reduction gears, and any drive where sound pressure is measured will show the gain. The same logic applies to gears that run at high load with a thin oil film, because a controlled flank surface keeps the film intact.

It does not pay off on slow, lightly loaded drives. A hand-fed gear in a packaging line at 30 rpm will not reveal a 10 μm profile improvement. Soft gears that never see heat treatment are usually cut to final size and shipped without a separate finishing operation.

Thin-walled gears are a real boundary case. A ring gear with a 4 mm wall can deflect during clamping and spring back after the pass, so the flank that measured well on the machine is out of tolerance in the fixture. We measure on the part, not on the arbor, and adjust the clamping pressure before committing to a process.

Internal gears and gears with shoulders near the face are harder to reach. The wheel or honing ring needs clearance to exit the tooth space. If the drawing has no relief groove, the process may need a smaller tool and more passes, which adds cost without adding accuracy.

Microgeometry is the small departure from a perfect involute: profile crowning, lead crowning, end relief, and bias. These features are measured in micrometres, not millimetres, and they decide where the teeth touch under load. A perfectly straight flank under torque bends toward the edges and concentrates stress there.

Crowning moves contact away from the edges. A typical automotive gear carries 5 to 15 μm of profile crowning and a similar lead crown, tuned so the contact patch stays inside the face under the design torque. Too little crown and the edges load up. Too much and the contact patch shrinks, raising local pressure and noise.

Surface roughness sets the oil film. A flank ground to Ra 0.8–1.6 μm holds a stable film at moderate speed. Below Ra 0.2 μm the surface can be too smooth to retain oil in some low-speed cases, and the risk of scuffing rises. That is why honing targets a band, not a single minimum value.

The link to noise is direct. Transmission error is the difference between the ideal and actual rotation of the driven gear. Profile and lead errors add to that signal, and the mesh amplifies it into whine. Finishing that tightens the flank geometry reduces transmission error at the source, which is cheaper than adding a sound package later.

• 1
  Profile crowningKeeps contact off the tooth tips under load.
• 2
  Lead crowningKeeps contact inside the face width.
• 3
  End reliefSoftens the entry and exit of the mesh.
• 4
  Roughness bandRa 0.2–1.6 μm depending on speed and load.

The clearest trend is closed-loop correction. A gear measuring center reports profile, lead, and pitch to the grinding machine, and the machine adjusts the next part automatically. On a stable process this cuts scrap during a run change and removes the manual touch-off that used to eat the first three parts.

The second trend is dry and minimum-quantity lubrication grinding. Coolant handling is a large share of the cost and the environmental load in a gear shop. Dry grinding with air or MQL needs a wheel that stays sharp at temperature and a machine rigid enough to hold geometry without the cooling cushion. It is used today on smaller modules and is spreading.

Honing has moved from a correction step to a controlled finishing step. Modern honing rings and force control let the process set microgeometry directly, not just roughness. That shortens the chain: grind close, hone to final, skip a second grinding pass.

Tooling has shifted toward CBN and ceramic wheels with defined dressing intervals. The gain is repeatability across a batch. When the wheel wears predictably, the correction model stays valid, and the operator stops chasing the first part of every lot.

On the metrology side, in-process and near-line measurement are replacing the gauge room. A flank measured minutes after the pass can be fed back before the batch is finished. For EV gears with tight noise targets, that feedback loop is now the difference between passing and reworking a lot.