Gear Parts Processing: How Teeth Are Actually Cut

• 1
  What the tooth profile decides before you cut metal
• 2
  Hobbing, shaping, and grinding compared
• 3
  Why heat treatment moves the teeth you just cut
• 4
  What to measure and when to measure it
• 5
  Material choice for gear parts processing
• 6
  Which gear cutting method fits your part

A gear is a torque converter with a fixed ratio. The tooth flank does the work, so almost every processing decision traces back to the profile. Involute teeth roll rather than slide at the pitch point, which keeps wear low and makes center-distance error tolerable. That is why the involute has held for two centuries.

Module and pressure angle set the whole geometry. A 20° pressure angle carries higher load per tooth than 14.5°, but it pushes radial force into the bearings. For a gearbox input stage, 20° is the default. For a lightly loaded instrument gear, 14.5° runs quieter because the contact ratio is higher.

Helical teeth spread contact across more than one tooth at a time. The result is smoother transfer and less noise, at the cost of an axial thrust load you must absorb somewhere. A 15° helix angle is a common compromise. Straight spur teeth are cheaper to cut and need no thrust bearing, so they still dominate low-speed drives.

Before any tooth is cut, the blank has to be right. Face runout over 0.02 mm will show up as a once-per-revolution noise, no matter how accurate the cutting was. Turn both faces and the bore in one setup, then mark the datum face. That single habit prevents most assembly complaints.

• 1
  ModulePitch diameter divided by tooth count. Sets tooth size.
• 2
  Pressure angle20° for load, 14.5° for quiet, lightly loaded gears.
• 3
  Helix angleTrades smoothness for axial thrust.
• 4
  Blank runoutKeep under 0.02 mm or accept once-per-rev noise.

Hobbing is the workhorse. A rotating hob feeds across the blank while the blank rotates in sync, generating the involute as a continuous process. It cuts spur and helical gears, it is fast, and it holds a lead accuracy that shaping struggles to match. For anything above roughly 20 teeth in volume, hobbing wins on cycle time.

Gear shaping uses a pinion-shaped cutter that reciprocates along the tooth. The cutter and blank mesh like a real gear pair. This is the method for internal teeth, for shoulders and flanges that block a hob from reaching the root, and for cluster gears where two gears sit close together. Shaping is slower, and the cutter is expensive, but it reaches places a hob cannot.

Grinding comes after heat treatment, not instead of it. Hardened teeth at 58–62 HRC cannot be cut, only ground. Profile grinding corrects distortion and lifts the accuracy grade, but it costs time and can leave grinding burn if the wheel is too aggressive. Use it on the gears that carry real load or set the timing of a system.

For one-offs and low volume, wire EDM and CNC milling can produce a usable tooth form without a dedicated cutter. The geometry is approximated rather than generated, so accuracy is lower, but the lead time is days instead of weeks. We use this route often for prototype gear parts processing before a hob is ordered.

• 1
  HobbingBest for external spur and helical gears in volume.
• 2
  ShapingInternal teeth, shoulders, and cluster gears.
• 3
  GrindingPost-hardening correction and higher accuracy grade.
• 4
  EDM or millingPrototypes and one-offs where cutter lead time hurts.

Case carburizing at 58–62 HRC gives a gear the wear resistance and core toughness to survive millions of cycles. It also distorts the part. Quenching is not uniform, so the bore grows, the teeth shift, and the lead twists slightly. A gear that measured Grade 7 green can measure Grade 9 after hardening.

The fix is to leave stock. Cut the teeth undersize by 0.05–0.15 mm per flank, harden, then grind back to final size. The exact allowance depends on module and section thickness, so the heat treater and the grinder need to agree on it before the first blank is turned. Guessing here is how gears end up scrapped.

Geometry matters as much as the process. A gear with a thick hub on one side and a thin web on the other will warp toward the thin side. Symmetric sections cool evenly. Where symmetry is impossible, a stress-relief pass before finish cutting removes the residual stress that would otherwise release during quenching.

Nitriding is the alternative when distortion must stay low. It runs at 500–550 °C, well below the transformation temperature, so the part barely moves. The case is thinner and the core stays soft, which suits lightly loaded gears, splines, and any part where grinding after hardening is impractical.

• 1
  Carburize and grind58–62 HRC, leave 0.05–0.15 mm per flank.
• 2
  NitrideLow distortion, thinner case, no post-grind needed.
• 3
  Symmetric sectionsEven cooling, less warp.
• 4
  Stress reliefBefore finish cutting when geometry is one-sided.

A gear is judged by its action, not its appearance. The four numbers that matter are profile error, lead error, pitch error, and runout. Profile and lead come off a gear measuring center. Pitch and runout can be checked on a CMM or with dedicated fixtures. Everything else is commentary.

Measure after heat treatment and after grinding, not only on the green cut. A green part that passes tells you the cutting was right. It says nothing about what the furnace did. If you only inspect once, inspect the finished part.

For prototypes, a functional check often beats a full inspection report. Mount the gear pair in its housing, apply the rated torque, and listen. A whine that rises with speed points to lead error. A knock at one position per revolution points to runout. Both are faster to diagnose on a running pair than on a chart.

We keep raw material certificates, in-process records, and final inspection reports for every gear run. Reports go out on request with the shipment. That paper trail matters when a customer needs to trace a lot back to a heat number.

• 1
  Profile and leadGear measuring center, after grinding.
• 2
  Pitch and runoutCMM or dedicated fixture.
• 3
  Functional checkRun the pair under load and listen.
• 4
  RecordsMaterial certs and inspection reports on request.

Through-hardened carbon steel such as 1045 is the low-cost baseline for lightly loaded gears. It machines cleanly, takes a decent finish, and hardens to around 50 HRC if needed. It is the right answer for a hand-cranked mechanism and the wrong answer for a gearbox that runs eight hours a day.

Alloy steels like 4140 and 4340 carry higher core strength after quenching, which is what stops a tooth from bending under shock load. They are the standard choice for automotive and industrial drive gears. 4340 is the tougher of the two and holds up better where the load reverses.

Stainless grades 303 and 17-4PH appear when corrosion matters. 303 machines easily but cannot be hardened much, so it suits lightly loaded gears in wet or washdown environments. 17-4PH can be aged to roughly 40 HRC and still resist rust, which makes it useful for food and medical equipment.

Plastics deserve a mention. POM and PA gears run quiet, need no lubrication, and cost little in volume. They cannot take high torque or high temperature, so keep them for instrument drives, toys, and light conveyors. Above roughly 1 N·m of continuous torque, switch back to metal.

• 1
  1045Low-cost baseline, light loads.
• 2
  4140 and 4340Shock and reversing loads.
• 3
  303 and 17-4PHCorrosion resistance, 17-4PH ages to about 40 HRC.
• 4
  POM and PAQuiet, dry-running, low torque only.