Surface grinding is the operation that turns a machined or heat-treated blank into a flat, parallel, dimensionally precise part — and it is also the line on a quote that estimators most often guess at. Milling and turning have intuitive cost models built around spindle hours; grinding does not. The surface grinding cost of a part is driven by how much stock you have to remove, how fine a finish the drawing demands, and how hard the material is — three things that have almost nothing to do with the part's footprint and everything to do with how many passes the wheel has to make.

Get those three wrong and the line is either padded enough to lose the job or thin enough to grind margin away one slow pass at a time. This guide breaks grinding into the inputs you can actually measure — stock removal, downfeed per pass, finish, hardness, dressing and setup — and shows how to turn them into a grind time and a defensible number, instead of a round figure tacked onto the back of a machining quote.

Why grinding doesn't price like milling

A milled feature costs roughly what the cutter takes to make it: material removal rate is high, and one or two passes do the job. Surface grinding removes material in tiny increments — hundredths of a millimetre per pass — because the wheel is finishing, not hogging. The cost is set by the number of passes, and the number of passes is set by how much stock sits above the finished dimension, divided by how little you dare take per pass.

That inverts the usual machining intuition. A large, simple plate that needs 0.4 mm ground off both faces can cost far more to grind than a small, fiddly part that needs a light 0.05 mm clean-up — even though the small part took longer to mill. And because grinding is usually a finishing step on already-machined or hardened stock, its cost rides on decisions made upstream: how much grind stock was left on, and whether the part was hardened (which is often why it's on the grinder at all). The companion build-up in how to price CNC machined parts treats grinding as one cost block among several — this article zooms into that block.

The five drivers of a grinding line

Five inputs fix the cost of grinding a surface. Estimate each and the line builds itself.

• Stock to remove — total material above the finished dimension, per surface, in millimetres. The single biggest driver of pass count.
• Downfeed per pass — how much the wheel drops each pass. Roughing passes take more; finishing passes take a fraction; spark-out passes take none but still cost time.
• Surface area and geometry — how far the table and wheel have to traverse to cover the face each pass. Bigger area means longer passes.
• Material hardness and grade — hardened tool steel, stainless and carbide grind slower and dress the wheel faster than soft steel or aluminium.
• Finish and tolerance — a fine Ra or a tight flatness/parallelism callout forces lighter passes, more spark-out, and a slower, more careful setup.

Add the grind time, the dressing time and the setup, multiply by the grinder's hourly rate, and you have the cost of the line.

Estimating grind time from stock removal

You don't need a CAM simulation to quote grinding. A pass-count model gets you within range:

1. Stock per surface (mm) ÷ downfeed per pass (mm) = number of cutting passes.
2. Add spark-out passes (typically 2–4) — these take no stock but clean up wheel deflection for finish and flatness.
3. Time per pass = the seconds to traverse the surface once, set by area and table feed.
4. Grind time = (cutting passes + spark-out passes) × time per pass.
5. Add dressing time (the wheel is trued periodically) and setup time (loading, indicating in, first-piece check).

Typical orientation values: roughing downfeed 0.02–0.03 mm/pass, finishing 0.005–0.01 mm/pass, spark-out 0.000 mm. Harder material and finer finish push you toward the low end of downfeed and the high end of spark-out — which is exactly how a tighter drawing turns into more passes and more cost.

A worked example: hardened steel plate

Take a hardened tool-steel plate, 200 × 120 mm, with 0.30 mm of grind stock left on one face, finished to a moderate Ra. The grinder runs at €75/hr (grinding rates sit at the top of the shop's spread — see machine shop hourly rate calculation for why a grinder hour costs more than a mill hour). Roughing at 0.02 mm/pass, finishing at 0.008 mm/pass, ~12 s per pass over this area:

Step	Calculation	Passes / time
Roughing (0.25 mm @ 0.02)	0.25 ÷ 0.02	~13 passes
Finishing (0.05 mm @ 0.008)	0.05 ÷ 0.008	~6 passes
Spark-out	fixed	3 passes
Grind time	22 passes × 12 s	~4.4 min
Dressing	per job	~1.5 min
Setup (loaded over batch)	15 min ÷ 10 parts	~1.5 min/part
Total per part	~7.4 min @ €75/hr	≈ €9.25

Now change one input. Tighten the finish so finishing runs at 0.004 mm/pass and add two spark-out passes: the pass count climbs, grind time pushes past 6 minutes, and the line moves toward €12 — a third more, from one tolerance callout. Double the stock left on to 0.6 mm and the roughing passes alone double. The arithmetic is simple; the inputs are where the money is.

What the drawing tells you about grinding cost

The grinder doesn't read the part — it reads the callouts. Three drawing details move a grinding line more than the part size does:

• Surface finish (Ra). A general ground finish is quick. A specified Ra below ~0.4 µm means lighter finishing passes and more spark-out, and below that you may be into lapping or honing territory — a different operation entirely.
• Flatness and parallelism. A tight GD&T callout on a face means slower, more careful grinding and sometimes a re-grind after stress relief. Reading what those callouts cost is its own skill — the GD&T and tolerances guide for estimators covers how a frame on the drawing becomes time on the machine.
• Hardness / heat-treat note. If the part is hardened, grinding is mandatory, not optional, and the wheel wears faster — more dressing, slower passes. The hardness note is a signal that a grinding line belongs on the quote at all.

None of this shows up if you price grinding as a flat percentage of the machining cost. All of it shows up in pass count.

Common mistakes that misprice a grinding line

• No grind stock budgeted upstream. If the milling step leaves nothing to grind, the part can't hit finish; if it leaves too much, the grinder eats hours. The grind stock is a decision, not an accident.
• Ignoring spark-out and dressing. They produce no stock removal, so they're easy to drop — and they're a real share of a short grinding cycle.
• Flat-rating every face the same. Grinding cost scales with area and stock, not with part count. A big face and a small face are not the same line.
• Pricing grinding as a markup on machining. Grinding inverts machining's cost logic; a percentage add-on overprices small parts and underprices large ones. Itemise it instead — the manufacturing quote template shows where a finishing line belongs.
• Forgetting the calendar. If grinding follows hardening, both add lead time, not just cost — fold the queue into the date you promise, the way manufacturing lead-time estimation lays out.

From a drawing to a costed grinding line

The slow, error-prone part of quoting ground parts is the reading: spotting the finish callout, the flatness frame, the hardness note, and judging how much stock sits above the finished dimension — all before you touch a pass-count model. Under deadline pressure that reading gets skipped, and the grinding line becomes a round number bolted onto a machining quote.

QuoteBuddy reads the technical drawing and surfaces what a grinding line needs — material and hardness, surface-finish and flatness callouts, the dimensions that imply how much stock has to come off — so the estimator works from a full picture instead of a glance. The estimating engine then builds the grinding line from your own grinder rate, downfeed and spark-out assumptions, dressing and setup times, and your target margin, the same way every time — so two estimators quoting the same part land on the same number.

Start a 30-day trial and run a few real hardened-and-ground drawings through it — from upload to a costed, itemised quote. See whether the grinding line it builds matches what you'd have priced by hand, and where a "10% for finishing" shortcut was quietly costing you on every plate.