Unit price hides the real difference. A line-by-line comparison of material yield, machining hours, die amortisation and fatigue life, and the batch size where the two routes cross over.
Photo: Parchaco / CC0 / Wikimedia Commons
Most gear quotations arrive as one number per piece, and most buying decisions are then made on that number. For a forged gear and a gear cut from bar, that number is misleading. The two routes consume different amounts of steel, different machine hours, different tooling and different inspection effort, and they fail differently after 50,000 kilometres. A fair comparison has to run from raw bar to a gear that passes the noise audit, not from loading dock to loading dock.
The categories are also blurrier than the words suggest. A gear sold as “forged” may be a near-net forging that needs only tooth cutting and grinding, or a rough forged block that still has to be turned, bored and hobbed from nothing. A gear sold as “cut” may come from a sawn billet, a hot-rolled bar, or a pre-turned blank somebody else supplied. Ask which of these four a quotation refers to before any comparison begins, because the cost gap inside one category is often larger than the gap between the two.
Near-net forging starts with a die cut to the shape of the finished part, less a thin machining allowance. The billet is heated, upset and closed-die forged so that the metal flows around the bore and web instead of being cut out of them. The result is a part that already has a bore, a web and often a pre-formed tooth space, with 1 to 2 mm of stock left for finishing. Grain flow follows the contour of the part, which is the physical property the whole argument rests on.
Cutting from bar is simpler to describe: saw, turn, bore, hob or shape, then heat treat and grind. A large share of the original bar leaves as swarf — typically 40 to 70 percent of the input weight for a shaft gear with a wide web. Machine time is longer because every surface is generated from solid, and interrupted cuts through mill scale shorten tool life. What the route buys in return is flexibility: any quantity from fifty pieces, no die investment, and geometry changes that cost a drawing revision rather than a new die.
There is a third route that buyers often overlook, and it is frequently the best compromise: a hot-forged blocker with generous allowance, normalised and then fully machined. It costs less than near-net forging in tooling and tolerance, keeps the continuous grain flow through the tooth root, and tolerates drawing changes better. For batches between roughly 500 and 3,000 pieces it is often the only route that is both affordable and metallurgically sound.
The table below sets out typical values for a passenger-car shaft gear in 20CrMnTi, module 3, case-carburised and ground. Ranges are wide on purpose: the point is the direction and scale of each difference, which holds across most gear families.
| Item | Near-net forged | Cut from bar |
|---|---|---|
| Material utilisation | 85–95% | 30–60% |
| Machining time per piece | 40–60% of the cut route | baseline |
| Grain flow at tooth root | continuous, follows contour | cut through, ends at surface |
| Bending fatigue life, relative | ≈1.2× | 1.0× |
| Forging die investment | USD 8,000–25,000 per die | none |
| Economic batch size | 3,000 pcs and above | from 50 pcs |
| Lead time, first lot | 10–14 weeks | 6–9 weeks |
| Unit price at 500 pcs | 15–25% higher | lower |
| Unit price at 20,000 pcs | 8–15% lower | higher |
| Dimensional consistency | uniform, low scatter | varies with bar lot |
Build the comparison as a cost per thousand pieces and the crossover becomes visible. Take a 1.6 kg finished gear. On the cut route the input bar is about 3.1 kg; on the near-net route it is about 1.75 kg. Steel is the same price per kilogram, so the material line alone moves by roughly 40 percent before any machining is counted. Roughing then disappears almost entirely: the forged part arrives with a bore and a web, so the CNC lathe takes one finishing pass instead of three.
| Cost line | Forged, near-net | Cut from bar |
|---|---|---|
| Steel input, kg | 1,750 | 3,100 |
| Material cost index | 56 | 100 |
| Rough machining hours | 60 | 210 |
| Tooth cutting hours | 90 | 95 |
| Heat treatment | equal | equal |
| Grinding hours | 70 | 78 |
| Die amortisation | 190 | 0 |
| Inspection and documentation | 40 | 45 |
| Total index | 96 | 128 |
Read the total line as a shape rather than a precise figure. Below about 2,000 pieces the die amortisation dominates and the forged route is more expensive per piece, typically by 15 to 25 percent at 500 pieces. Between 2,000 and 3,000 pieces the two lines meet. Above that the material and roughing savings compound, and the forged route lands 8 to 15 percent cheaper at 20,000 pieces — before the fatigue benefit is priced at all.
Two adjustments are often missed. Freight works in favour of forging, because shipping 1.75 kg of steel instead of 3.1 kg cuts import weight by more than 40 percent; on air freight this alone can outweigh a year of die amortisation. Scrap credit works the other way: a well-organised cutting shop sells its swarf, and that revenue quietly subsidises the cut route by a few percent. Both belong in the model, and both are routinely left out.
Forging is not a universal answer, and pretending otherwise damages credibility with engineers who know better. Four situations favour the cut route:
The published advantage of forging — roughly six to ten times the impact resistance of cut gears in some studies, and about 20 percent more fatigue life — belongs to the process, not to the word. It exists only when grain flow is genuinely continuous around the root fillet and when case depth at the root is correct. A forging that is later over-ground at the root, or heat treated so that the case is shallow exactly where bending stress peaks, delivers none of the benefit while carrying all of the cost.
Three checks settle it in an afternoon. Section a sacrificial part through a tooth, polish and macro-etch it: the fibres should run along the flank and turn continuously around the root. Map case depth at three positions on three teeth and look at the scatter, not the average. And ask for the dimensional capability study rather than the tolerance table, because a forging that wanders batch to batch forces the machinist to protect against the worst case, which quietly deletes the saving.
For a programme above 3,000 pieces a year, with a stable drawing and a root-bending or fatigue-critical duty, specify near-net forging and audit the grain flow. Between 500 and 3,000 pieces, specify a normalised blocker forging plus full machining: it keeps the metallurgy without the die cost. Below 500 pieces, or where the drawing is still moving, cut from bar and revisit the decision once the design freezes. In every case, compare total cost per accepted piece — including scrap, freight, die amortisation and the inspection needed to prove the claim — rather than price per piece at the dock.
