CNC machining vs gear hobbing for prototype gears

I have spent the last 18 years sourcing prototype gears, and the question I hear most is simple: should I cut the teeth on a CNC mill or run them through a gear hobbing machine? The honest answer is that it depends on quantity, tolerance, and how fast you need the parts on your desk. Let me walk through what I have learned the hard way, so you can avoid the mistakes I made.

What each process actually does to the gear teeth

CNC milling cuts each gear tooth one at a time using a ball or end mill. The machine traces the involute profile in small step-overs, and the surface finish depends on the step-over distance. A smaller step-over gives a smoother flank but takes longer. I have seen operators run 0.05 mm step-overs for a 24-tooth module 2 gear and spend 40 minutes per part. You can hold plus or minus 0.025 mm on tooth-to-tooth spacing if the machine is in good shape and the fixture is rigid.

Gear hobbing uses a rotating hob that meshes with the blank as it spins. The hob cuts several teeth at once in a continuous motion. A standard hobbing machine can finish that same module 2 gear in under 4 minutes. The flank finish is uniform because the cutting action is generating, not point-to-point tracing. You get a consistent pressure angle and involute shape, which matters if the gear runs in a transmission at high speed.

The difference in cycle time is the first thing buyers notice. But cycle time alone does not decide the question. The real trade-offs are tolerance flexibility, material selection, setup cost, and what happens when you revise the design.

Cost per part: where the crossover point sits

For 1 to 5 prototype gears, CNC milling wins. You load a blank, pick a stock end mill, and hit cycle start. No special tooling, no hob to order. The shop charges $80 to $180 per gear depending on material and complexity. I paid $145 each for a batch of three 17-4PH spur gears last year, and the parts met my drawing within plus or minus 0.02 mm.

Gear hobbing needs a hob that matches your module, pressure angle, and tooth count range. A single DIN 3971 class AA hob runs $300 to $900. The machine setup adds another $150 to $250 in fixturing and first-article alignment. If you are making 5 parts, you spend $600 on tooling and setup to save 30 minutes of cycle time per part. The math does not work. At 50 parts the hob pays for itself. At 200 parts the unit cost is half of CNC milling.

Here is the crossover I use as a rule of thumb: below 10 parts, CNC mill. Above 50 parts, gear hobbing. Between 10 and 50, get both quotes and check lead time. I have seen shops quote hobbing at 20 parts because they already had the hob on the shelf from a previous job. That changes the math completely.

Tolerance and tooth quality: what you can actually hold

CNC milling can hit DIN 8 to DIN 9 gear quality if the programmer knows what they are doing. The problem is that the involute profile is only as good as the CAM post-processor and the tool wear. After cutting 15 gears in 4140 steel at 32 HRC, the end mill develops flank wear that shifts the tooth thickness by 0.03 to 0.05 mm. I caught this on a transmission shaft gear in 2021. The first three parts were perfect. Part 12 was 0.04 mm oversize on tooth thickness, and the gear started making noise in the test rig.

Gear hobbing holds DIN 7 on a decent machine with a fresh hob. DIN 6 is possible with a class AA hob and temperature-controlled environment. The generating cutting action produces a more consistent involute because the hob geometry defines the profile, not a tool path that drifts as the cutter wears. For a gear that runs at 3,000 RPM or higher, tooth quality matters more than it does for a low-speed actuator gear.

I always ask for a single-flank or dual-flank gear inspection report on prototype gears. If the shop cannot produce one, I walk away. You can learn more about gear quality requirements in our gear hobbing service page.

Material flexibility: when CNC is the only option

This is where CNC milling has a real edge. You can machine a gear from almost any material: pre-hardened 4140 at 28 HRC, 17-4PH H1025, 316L stainless, titanium Ti-6Al-4V, even tool steel M2 at 60 HRC if you have a good carbide end mill. The milling cutter does not care about tooth count or pressure angle because the CAM generates the path.

Gear hobs are material-specific to a degree. A hob designed for soft steel will dull fast on hardened 4140. Coated hobs (TiAlN or AlCrN) can cut harder materials but cost more and still wear at 35 to 40 HRC. I tried hobbing a batch of 17-4PH H1150 gears and lost 15% of hob life compared to annealed 316L. For exotic alloys, CNC is usually the safer bet at prototype stage. You can check what material options are available on our titanium Ti-6Al-4V material page and stainless steel 316L material page.

One more thing: internal splines and internal gears. CNC can cut internal teeth with a slotting cutter or wire EDM. Gear hobbing only works on external teeth unless you use a special spline hobbing setup, which most shops do not have. If your prototype has an internal spline, CNC or EDM is your path.

Design changes: the prototype reality

Prototypes get revised. I have never seen a prototype gear go to production without at least one tooth count change or face width adjustment. When you change tooth count on a CNC-milled gear, you update the CAM file and hit run. Twenty minutes of programming, same tool, same setup.

Change tooth count on a hobbed gear and you might need a different hob. If the new tooth count is outside the range of the existing hob, you buy a new one. Two week lead time from the hob manufacturer. I sat through that delay on a planetary gear set for an aerospace actuator in 2019. We changed the sun gear from 20 teeth to 22 teeth and waited 11 days for the new hob. The program slipped by a full sprint.

For that reason alone, I lean toward CNC for the first prototype iteration. Once the design is frozen and you are moving to pilot production, gear hobbing becomes the better process. You can explore how we handle this transition on our how we work page.

Lead time comparison for a typical prototype run

Here is what I see from Chinese shops for a standard module 1.5 spur gear, 24 teeth, 20mm face width, 316L stainless:

• CNC milling, 5 parts: 3 to 5 days after drawing approval. No special tooling. Setup is 30 minutes per part.
• Gear hobbing, 5 parts: 10 to 14 days. The hob may need ordering. Setup and first-article take 3 to 4 hours. Once running, cycle time is under 3 minutes per part.
• CNC milling, 50 parts: 8 to 12 days. Same setup, but the operator runs them unattended overnight.
• Gear hobbing, 50 parts: 12 to 16 days including hob procurement. After setup, the machine runs 50 parts in under 3 hours.

The hobbing lead time looks longer because of tool procurement. But if the shop stocks your module hob, it drops to 5 to 7 days for any quantity. Always ask if they have the hob on hand before committing. You can read about typical questions to ask suppliers on our FAQ page.

When to pick which process: my decision framework

After doing this for almost two decades, here is how I decide:

Pick CNC milling for prototype gears when: you need 1 to 10 parts, the design is still changing, the material is exotic or pre-hardened, the gear has internal teeth or unusual geometry, or you need the parts in under a week. The process flexibility is worth the higher unit cost at low volume.

Pick gear hobbing for prototype gears when: you need 20+ parts, the tooth profile is standard involute, the material is machinable steel under 35 HRC, the gear runs at high speed where tooth quality matters, and you can wait for hob procurement. The per-part cost drops fast and the tooth quality is more consistent.

The transition from prototype to production is where the decision gets locked in. Once you commit to hobbing for production, your design flexibility drops because every tooth count change means a new hob. I recommend CNC for iterations 1 and 2, then move to hobbing for the production design freeze. This approach has saved my clients an average of 35% on prototype-to-production cost across the last 12 projects I managed.

If you are sourcing gears for automotive applications, the standards are tighter. IATF 16949 suppliers need PPAP documentation and Cpk data. Prototype gears for automotive validation runs often start on CNC and transition to hobbing at the B-sample stage. For aerospace applications, AS9100 traceability requirements mean you document every step of the process change, including the CNC-to-hobbing transition.

I have seen buyers pick gear hobbing for a 3-piece prototype because a sales rep told them it was the professional way to make gears. They spent $700 on tooling to save $60 in cycle time. Do not do that. Match the process to the quantity, not to the marketing.

The process question comes down to quantity, tolerance, material, and how frozen your design is. If you are still iterating, CNC gives you speed and flexibility. If you are heading to production, hobbing gives you consistency and lower unit cost. Get both quotes, compare them honestly, and pick the one that fits your current phase. You can learn more about our CNC machining capabilities and gear hobbing services to see what we can do for your prototype run.