Grinding vs Hard Turning Cost: Which Wins for Hardened Parts
September 16, 2026
Still quoting parts like this by hand? Try QuoteBuddy free — from drawing to quote in minutesFree for 30 days · Unlimited quotes · No credit card required
Every shop that hardens steel eventually faces the same routing decision: the part comes back from heat treatment at 60 HRC needing finished bearing seats, and the traveller can send it either to the cylindrical grinder or back to a lathe with CBN inserts. Grinding vs hard turning cost is one of the few process choices where the cheaper option can be two to three times cheaper — or flatly impossible — depending on the tolerance block, the geometry, and the batch size, which makes it a decision worth pricing properly rather than defaulting to whatever the shop has always done.
This guide is decision content, not process advocacy. It lays out what each route actually costs per part — cycle time, machine rate, tooling, setup — where hard turning's surface finish and roundness genuinely stop, the batch-size logic that flips the answer, and the cases where grinding is simply mandatory. By the end you should be able to look at a hardened-part drawing and know not just which process wins, but by how many euros.
What hard turning is, and when it enters the conversation
Hard turning is single-point turning of hardened steel — typically 55–65 HRC — on a rigid CNC lathe using CBN (cubic boron nitride) or, for less demanding work, ceramic inserts. It is not exotic anymore: a good-condition lathe with adequate rigidity, decent spindle accuracy, and the right insert grades will finish hardened bearing seats, seal diameters, and faces that a generation ago went to the grinder by default.
The economic appeal is structural. A lathe finishes multiple features in one chucking — several diameters, shoulder faces, chamfers, a groove — where a cylindrical grinder typically addresses one or two diameters per setup. Hard turning usually runs dry, removes material several times faster than grinding for equivalent stock, and lives on a machine you already own and already know how to quote, using the same logic as any turning job. The catch is the finish line: below certain roundness, finish, and geometry thresholds, the grinder is not the expensive option — it is the only option, and knowing where that line sits is the whole game.
The cost blocks: cycle, rate, tooling, setup
Both routes decompose into the same four blocks, and the comparison is honest only if all four are priced.
Cycle time. For typical finishing stock (0.15–0.4 mm on diameter), hard turning removes it in a single pass at cutting speeds of 120–200 m/min. Cylindrical grinding takes it in multiple sparks-out passes with wheel dressing between parts or batches. For a shaft with two or three ground features, hard turning cycle is commonly 40–70% shorter — and the gap widens when the part has faces and grooves the grinder would need extra plunges or a second setup for.
Machine rate. A production cylindrical grinder is usually the more expensive asset per hour once depreciation, dressing infrastructure, filtration, and a skilled operator are burdened in — typical 2026 calibration is €75–110/h for the grinder against €60–90/h for a rigid CNC lathe. Neither number is universal; build both from your own burdened machine rates, because the wrong blended rate will silently decide this comparison for you.
Tooling. This is where hard turning pays its toll. A CBN insert edge costs €25–60 against a few euros for a carbide edge, and edge life in hardened steel is measured in tens of parts, not hundreds — interrupted cuts (keyways, cross-holes, splines) can halve it again. Grinding wheels are cheap per part in comparison: a wheel lasts thousands of parts, and even with dressing consumables the per-part tooling figure rarely exceeds €0.30–0.80. On tooling alone, grinding wins; the question is whether cycle and setup give it back.
Setup. A lathe setup for a hard-turned finishing operation — chuck or collet, two tools, prove-out — typically runs 30–60 minutes. A cylindrical grinder setup with wheel selection, dressing, and gauging setup commonly runs 45–90 minutes, and a part needing OD and face grinding may need two setups where the lathe needed one. On small batches this block alone can decide the comparison.
Side-by-side: the numbers that matter
Calibration ranges for 2026 job-shop work on hardened steel (55–62 HRC), ordinary shaft-type parts. Your machines and specs move every row — treat these as the frame for your own numbers, not a substitute for them.
| Factor | Hard turning (CBN) | Cylindrical grinding |
|---|---|---|
| Machine rate, burdened | €60–90 / h | €75–110 / h |
| Cycle, 2-feature shaft seat set | 3–6 min | 6–14 min (often 2 setups) |
| Tooling cost per part | €0.80–3.00 (CBN edges) | €0.30–0.80 (wheel + dress) |
| Setup per batch | 30–60 min | 45–90 min per setup |
| Achievable Ra | 0.2–0.4 µm (good setup) | 0.05–0.2 µm |
| Achievable roundness | 1–3 µm | 0.2–1 µm |
| Practical tolerance class | IT5–IT6 | IT3–IT5 |
| Coolant | Usually dry | Flood + filtration |
Read the table from the bottom up. If the drawing demands Ra 0.1 or roundness under a micron, the cost rows are irrelevant — the part grinds. If the drawing sits inside hard turning's window (Ra ≥ 0.3, roundness ≥ 2 µm, IT6), the cost rows take over, and they mostly favour the lathe.
Worked example: a hardened shaft, both routes in euros
A gear shaft, 42CrMo4 hardened to 58 HRC, two bearing seats (Ø35k5, Ra 0.4) plus shoulder faces, batch of 60. Finishing stock 0.3 mm on diameter, left before heat treatment — the stock allowance and distortion behaviour that make this possible are part of quoting the heat treatment step properly.
Hard turning route (lathe at €75/h, one setup):
- Cycle: 4.0 min per part → €5.00
- CBN tooling: €45 per edge, 25 parts per edge, 2 edges in cut → €1.80 per part
- Setup: 45 min = €56.25 ÷ 60 parts → €0.94
- Cost per part ≈ €7.74; batch ≈ €464
Grinding route (grinder at €90/h, OD + face in two setups):
- Cycle: 8.5 min per part → €12.75
- Wheel and dressing: → €0.55 per part
- Setup: 2 × 70 min = €210 ÷ 60 parts → €3.50
- Cost per part ≈ €16.80; batch ≈ €1,008
Hard turning takes this part at less than half the grinding cost — €544 saved on one modest batch — and the drawing permits it: k5 with Ra 0.4 sits inside the lathe's window on a rigid machine. Move one spec — Ra 0.2, roundness 0.8 µm, or a customer note reading "grind only" — and the right-hand column stops being expensive and becomes the price of the job. That is why this comparison belongs in the quote stage, not on the shop floor after the order is won.
Where hard turning stops: finish, roundness, geometry
The limits are physical, and pretending otherwise produces scrap at 60 HRC prices.
Surface finish. A well-set-up hard turning operation reliably delivers Ra 0.3–0.4 µm, and on its best day touches 0.2. Below that — seal running surfaces at Ra 0.1, superfinish-prep at 0.05 — the single-point process runs out of road, and the feed marks that are invisible at 0.4 become functional at 0.1. The cost consequences of finish callouts deserve their own respect regardless of process, as covered in surface finish cost impact.
Roundness and size control. Hard turning inherits the lathe's spindle error and thermal drift; 1–3 µm roundness is realistic on a rigid machine in good condition. Precision bearing seats specified under a micron, or size tolerances at IT4 and below, belong on the grinder, whose geometry-generating mechanics are simply better at it.
Geometry and stability. Thin-walled parts deflect under single-point cutting force where a grinding wheel's distributed contact is gentler. Long slender shafts (L/D above roughly 8–10 without steadies) chatter. Heavily interrupted surfaces eat CBN edges. And on safety-critical fatigue parts, a worn CBN edge can leave a rehardened "white layer" that some aerospace and bearing specs explicitly prohibit — a metallurgical, not economic, veto.
The paperwork veto. If the drawing or the customer's process spec says grind, it grinds. Process substitution on a controlled drawing is a concession request, not an estimator's shortcut.
Batch size flips the answer
The two routes carry their fixed and variable costs in opposite places, so quantity moves the crossover. Hard turning's setup is short but its tooling cost per part is high and constant; grinding's setup is long but its per-part tooling is nearly free. Using the worked-example numbers:
| Batch size | Hard turning, per part | Grinding, per part |
|---|---|---|
| 5 | €18.05 | €55.30 |
| 20 | €9.61 | €23.80 |
| 60 | €7.74 | €16.80 |
| 250 | €7.03 | €14.14 |
| 1,000 | €6.86 | €13.51 |
On this part the lathe wins at every quantity — the cycle gap is too large — but notice the shape: grinding improves faster with volume as its heavy setup amortises. On parts where the cycle gap is narrow (single plunge-ground diameter, no faces), high volumes and cheap wheel-life can genuinely hand the win back to the grinder, which is precisely how dedicated grinding cells earn their keep in production plants. For job-shop quantities — fives, twenties, sixties — the short-setup, one-chucking economics of hard turning dominate whenever the spec allows it. The full grinding-side arithmetic, including dressing cycles and sparks-out allowances, is built up in cylindrical grinding cost quoting.
Quoting the decision, not just the process
The practical estimating workflow is a three-question gate. First, does the tolerance block permit hard turning — finish, roundness, IT grade, white-layer or "grind" notes? If no, quote grinding and stop. Second, does the geometry permit it — rigidity, interruptions, L/D? Third, if both permit, price both routes with your own rates and pick deliberately, keeping the loser in the file: when the customer comes back doubling the quantity or tightening the seat tolerance, the re-quote is a lookup instead of a redo.
One further note on honesty with yourself: if your shop has no grinder and subcontracts grinding, the comparison is your hard-turning cost against the subcontractor's price plus transport and an extra week of lead time — which changes the arithmetic, usually in the lathe's favour, and belongs explicitly in the quote rather than absorbed as "logistics".
FAQ: hard turning vs grinding
Is hard turning cheaper than grinding? When the spec allows it, usually yes — commonly 30–60% cheaper per part for shaft-type work, because the lathe finishes several features in one short-setup chucking while the grinder needs longer cycles and often multiple setups. The saving survives CBN's expensive edges because cycle and setup dominate tooling on most parts. When the spec demands sub-micron roundness or Ra below 0.2 µm, the question is moot: it grinds.
What surface finish can hard turning achieve? A rigid machine, sharp CBN edge, and correct feed reliably produce Ra 0.3–0.4 µm; 0.2 is achievable but should be treated as the edge of the envelope, not a quoting default. Grinding routinely delivers Ra 0.05–0.2 µm. Quote hard turning at the finish you can hold every day, not the one the demo part achieved.
When is grinding mandatory? Roundness below roughly 1 µm, size tolerances at IT4 or tighter, finishes below Ra 0.2, thin-walled or slender parts that deflect under cutting force, specs that prohibit white layer on fatigue-critical surfaces, and any drawing or customer process spec that explicitly calls out grinding. In all of those cases hard turning is not the cheap option — it is the wrong one.
How much does CBN tooling cost? Budget €25–60 per usable cutting edge for quality CBN grades in 2026, with edge life in continuous cuts of hardened steel typically 15–40 parts depending on hardness, stock, and finish demands — and materially less on interrupted cuts. Per part that is €0.80–3.00, an order of magnitude above carbide turning but still small against the cycle-time saving on most jobs.
Does hard turning need a special lathe? It needs rigidity, good spindle condition, and stable thermals more than it needs a badge. Many shops hard turn successfully on quality standard CNC lathes; worn-out machines that turn soft steel acceptably will betray you at 60 HRC, where every micron of spindle error prints onto the part.
From hardened-part drawing to a priced decision
The grinding-versus-hard-turning call starts with reading: the tolerance block on the seat diameters, the Ra callouts, the roundness spec hiding in the GD&T, the "grind" note in the title block, the hardness callout that says this part comes back from heat treatment at all. Miss one of those under deadline pressure and the quote prices the wrong process — cheap and unmanufacturable, or safe and uncompetitive.
QuoteBuddy reads the technical drawing and surfaces exactly those inputs — diameters and their tolerance classes, surface finish callouts, geometric tolerances, material and hardness notes — so the estimator sees the decision instead of hunting for it. The deterministic cost engine then prices the routing from your own machine rates, cycle estimates, tooling costs, and setup times, the same way every time, so the hard-turn-or-grind call is made once, on numbers, and repeats consistently across every estimator in the shop.
Start a 30-day trial and run a hardened part through it — both routes, your rates — and see which process your margin actually prefers.