Gear Tooth Manufacturing: How Teeth Are Actually Cut, Ground and Finished
This page explains the main gear tooth manufacturing routes, what each one does to the blank, and where each process stops being the right choice. It is written for design and process engineers who need to pick a method before the drawing is frozen.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
What gear tooth manufacturing really changes
Gear tooth manufacturing is the set of operations that turn a round or near-round blank into a toothed part with a defined involute, a defined pitch, and a defined surface at the flank. Everything before the first cut decides how hard that job will be: material condition, blank concentricity, bore tolerance, and whether the part has a shoulder or a flange that the cutter has to clear.
The tooth itself is only one feature. A gear fails or passes on the relationship between the tooth flank, the bore, the mounting face, and the reference diameter used during inspection. If the bore is drilled after the teeth are cut, runout moves. If the teeth are cut after heat treatment, the cutter wears fast. Process order is a design decision, not a shop-floor habit.
Three families cover most work. Form cutting shapes the gap with a tool that matches the profile. Generating methods roll the tool and the blank together so the involute comes from the motion. Grinding removes a small amount of material from a hardened flank to correct heat-treat distortion and hit the final tolerance.
At GreatLight we see gear-like parts across prototyping and low-to-mid volume runs, often as one element of an assembly. That is the context for the rest of this page: how to reason about the process, not how to run a gear shop.
Hobbing and shaping: the two generating routes
Hobbing uses a rotating cutter with a helical thread that acts like a rack in motion. The hob and the blank rotate in a timed relationship, and the cutter feeds along the face width. One hob covers a range of tooth counts for a given module and pressure angle, so setup time is short and the cycle is continuous. It is the default for spur and helical external gears.
Hobbing cannot cut a tooth that has no clearance for the hob to run out. Internal gears, gears next to a shoulder, and cluster gears with a small gap between them are outside its reach. The hob also leaves a visible feed pattern on the flank, which matters if the gear runs against a mating part at high speed.
Shaping uses a pinion cutter that reciprocates and rotates with the blank. It reaches internal teeth, and it cuts close to a shoulder because the cutter only needs a small runout. The trade-off is cycle time. Shaping is slower than hobbing per tooth, and the cutter is more expensive to maintain. For a small batch of internal gears, that cost is usually still lower than a broach.
Both methods generate the involute from the relative motion, so the tooth form is as accurate as the machine timing and the cutter. A worn hob or a loose workholding setup shows up as profile error and lead error, not as a visibly bad tooth. That is why in-process checks on pitch and profile matter more than a visual look at the part.
- 1HobbingExternal spur and helical gears, high volume, one hob per module range.
- 2ShapingInternal teeth, shoulders, cluster gears, small batches.
- 3BothGenerating motion defines the involute; cutter condition defines the error.
Milling and form cutting when the geometry is awkward
Form milling cuts each tooth gap with a disc or end mill shaped to the profile. The dividing head indexes the blank between cuts. Accuracy depends on the number of teeth the cutter was designed for, so a single cutter is only exactly right for one tooth count and is approximate for others. It is a reasonable route for coarse pitch, low quantity, or a repair.
CNC milling of gear teeth on a 3-axis or 4-axis machine is a different case. With a ball or radius cutter and a CAM path, the flank can be approximated to whatever deviation the drawing allows. This suits large gears, odd profiles, sector gears, and prototypes where a hobbing cutter would be a special order. It is slow per tooth and the flank finish depends on stepover.
Five-axis work helps when the tooth is not straight or the part has features that need to be reached from two directions without re-fixturing. A single setup keeps the tooth-to-bore relationship tight. For a one-off or a small batch, that can beat a dedicated gear machine on total lead time.
The boundary is honest: for a standard spur gear in a run of thousands, form milling and CAM milling lose on cost and repeatability. For a large, unusual, or low-quantity gear, they are often the only practical route.
- 1Form millingCoarse pitch, low quantity, repairs; cutter is tooth-count specific.
- 2CAM millingLarge gears, sector gears, odd profiles, prototypes.
- 35-axisComplex parts, one setup, tight tooth-to-bore relationship.
Heat treatment, distortion, and why grinding exists
Most steel gears that carry load are case hardened or through hardened after cutting. The quench moves the part. Bore runout, face runout, and flank profile all shift by amounts that are large compared with a fine-pitch tolerance. Leaving grind stock of 0.05–0.15 mm per flank is common on parts that will be ground; the exact figure depends on section size and the hardening process.
Grinding removes that stock with a profiled wheel or a generating wheel. Form grinding uses a wheel dressed to the tooth profile and indexes tooth by tooth. Generating grinding rolls the wheel and the workpiece, which is faster for long runs and handles profile modification in the dress. Both correct the distortion from heat treatment rather than causing it.
Grinding is where the final tolerance and flank finish are set. It also introduces its own risks: burn, a soft spot from tempering, and cracks at the root if the wheel is too hard or the feed too aggressive. A nital etch check on a sample is a normal way to confirm the surface is sound.
Not every gear needs grinding. A soft gear running at low speed with a generous backlash and a loose tolerance can go straight from hobbing to deburring. Grinding adds cost, lead time, and a second setup. The decision should come from the function, not from a habit of specifying the tightest callout available.
Choosing a route from material, size, and quantity
Material sets the cutting speed and the tool life. Aluminum and brass cut cleanly and let you use higher spindle speeds and lighter fixtures. Stainless steels such as 303, 304, and 17-4PH work harden, so a light pass that rubs instead of cutting will ruin the flank and the cutter. Low-carbon steels like 1018 and 1045 cut well but need heat treatment before they can carry load.
Size sets the machine. Small gears with fine pitch need a machine with low runout at the work spindle and a fine feed. Large gears need a table that carries the part without deflection and a cutter with enough reach. On our floor the maximum processing size is 4,000 mm, and the large-travel machines run 4,000 × 400 × 150 mm. Those numbers define what fits, not what is economical.
Quantity decides the tooling. A special hob or a formed grinding wheel has a cost that has to be spread over the run. One-off parts favor CAM milling or a standard cutter. Runs in the thousands favor generating methods with a dedicated setup. Runs of 10,000 and up can justify a broach or a dedicated line.
There is no universal best process. There is a process that matches the tolerance, the material, the geometry, and the batch size at the same time. Change one of those four and the answer changes.
Gear tooth process comparison
Use this as a first filter, then confirm with a DFM review.
| Process | Typical use | Tolerance and finish | Main limit |
|---|---|---|---|
| Hobbing | External spur and helical gears | Ra 1.6–3.2 μm after cut | No internal teeth or tight shoulders |
| Shaping | Internal gears, shoulders, clusters | Ra 1.6–3.2 μm after cut | Slower cycle, costlier cutter |
| Form milling | Coarse pitch, repairs, low quantity | Ra 3.2 μm and up | Cutter correct for one tooth count |
| CAM milling | Large, odd, or prototype gears | Ra 1.6–3.2 μm with fine stepover | Slow per tooth, path deviation |
| Grinding | Hardened flanks, final tolerance | Ra 0.2–0.8 μm achievable | Burn and crack risk, extra setup |
| Broaching | Internal teeth, high volume | Tight pitch and profile | High tooling cost, one geometry |
Which route to pick
If the gear is external, standard, and runs in volume, hob it and grind only the flanks that carry load. If it is internal, has a shoulder, or is a one-off with unusual geometry, shape it or mill it on a 4-axis or 5-axis machine and skip the dedicated tooling. Match the process to the geometry and the batch, not to the tightest tolerance you can write on the drawing.
Gear tooth manufacturing questions
How much grind stock should I leave on a hardened gear?
For most case-hardened steel gears, 0.05–0.15 mm per flank is enough to clean up distortion from the quench. Deep-case or large-section parts can move more, so the figure should come from the heat treater and a trial part.
Too little stock leaves hard spots that the wheel cannot clean. Too much stock turns grinding into a roughing operation and raises burn risk. Both show up as scrap at final inspection.
Can you cut gear teeth after heat treatment?
Cutting a hardened flank with a hob or a form cutter is not practical in most cases; the tool wears out and the surface tears. The normal order is cut soft, harden, then grind the flanks.
If the part is nitrided rather than case hardened, the case is thin and the distortion is small. In that case a light grind or, for loose tolerances, no grind at all can be enough. Confirm this at the design stage rather than after hardening.
What causes a gear to run noisy even when the dimensions check out?
Pitch error, profile error, and lead error all change the contact pattern and produce noise without moving the measured diameter out of tolerance. A single-point size check will not catch them.
Check the tooth-to-tooth pitch variation, the profile trace, and the lead trace. Also check runout between the bore and the tooth flank. Noise usually comes from the relationship between features, not from the size of one feature.
Is 5-axis milling a replacement for hobbing?
No. For a standard external gear in any real quantity, hobbing is faster and more repeatable. Five-axis milling is the right answer when the geometry is unusual, the batch is small, or the part needs several features in one setup.
We run 16 simultaneous 5-axis machining centers, and they are used for exactly those cases: complex parts where re-fixturing would cost more accuracy than the CAM path costs in cycle time.
What information do you need to quote a gear part?
Send the drawing with module or diametral pitch, tooth count, pressure angle, helix angle, profile shift if any, material and heat treatment, and the tolerance on the bore and the tooth flank. A 3D model helps but does not replace the tooth data.
If the gear meshes with an existing part, send that part or its data too. The backlash and the center distance are set by the pair, not by one gear alone.
Do you handle small quantities?
Yes. There is no minimum order quantity, and runs can go from one prototype to 10,000+ parts. Quotation and a free DFM analysis come back within 12 hours, and production can start within 24 hours of approval.
For a single gear, expect a CAM-milled or shaped part rather than a dedicated hobbing setup. The DFM note will say which route we recommend and why.
Send the gear data, get a route and a price
Upload the drawing and the mating part data. We will come back with a DFM note that names the process, the tolerance we can hold, and the price.
12-hour quoteDFM analysis includedNo minimum order quantity
More from the shop floor
We publish setup notes, tooling trials and inspection data from the factory floor.
