CNC engine block processing
How a cast or billet block becomes a dimensioned, sealed, running part. We cover datums, bore geometry, deck flatness and the setups that decide whether the numbers hold. Written for engineers and buyers who have to approve a process, not just read about one.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
What the process actually removes
An engine block arrives as a casting or a billet with a few millimeters of stock on every functional surface. The job removes that stock until the block has datums, bores and sealing faces that another part can bolt to and seal against. Nothing here is decorative. Every cut either creates a reference, holds a fit, or opens a passage.
The sequence matters more than the spindle. Operators rough the block first, let it cool, then semi-finish and finish. Roughing a 20 kg aluminum casting moves several hundred grams of material and adds heat. If the finish passes run straight after roughing, the block is still growing and shrinking while the tool cuts. Bore roundness drifts.
Stock allowance sets the floor on cost. A casting with 1.5 mm of stock per face needs fewer roughing passes than a billet hogged from 150 mm plate, but a casting with sand inclusion or shifted cores forces extra cleanup passes on the same surfaces. We inspect the raw casting before quoting so the allowance is realistic.
Thin-wall blocks behave differently. Wall sections under 4 mm deflect under normal cutting force, so we reduce radial engagement and step down in smaller increments. The result is a longer cycle, not a scrap part. That trade is worth making on prototype blocks where wall thickness is still being tuned.
Setting datums before cutting anything
A block has no useful geometry until someone declares what it sits on. The first operation establishes the primary datum, usually the pan rail or a machined pad on the bottom face. Everything downstream is measured from that plane. If the first cut is off by 0.05 mm, the cam bore and the deck inherit that error.
Cylinder bore position ties directly to the crank centerline. A 0.02 mm shift at the crank bore becomes a measurable change in piston-to-wall clearance at the top of the stroke. On a 90 mm bore running 0.05 mm clearance, that shift eats 40 percent of the margin.
We use the same datum scheme in fixturing that we use in inspection. That sounds obvious. It gets skipped when a shop programs from the CAD model origin instead of the physical block, and the first article report then disagrees with the CMM report by a fixed offset.
Datum transfer between operations is where error stacks. Every time the block moves to a new fixture, the operator repeats the locating scheme. On a five-operation block, five setups can stack 0.01 mm of variation each. The total is what shows up on the deck, not any single setup.
Why 5-axis setups hold the tight features
Reaching a deck face and the crank bore in one setup removes the largest single error source: re-fixturing. A simultaneous 5-axis center tilts the tool and the table together, so a compound-angle oil gallery or a deck cut with draft can be reached without unclamping the block. Fewer clamps mean fewer chances to lose the datum.
Tool access drives the choice more than axis count does. A deep bore with a cross-passage needs a long, thin tool that deflects. Shortening the tool by tilting the head is often the difference between holding 0.010 mm roundness and chasing it. That is a rigidity decision, not a feature-count decision.
Not every block needs five axes. A straight four-cylinder block with open sides machines fine on a 3-axis mill with two or three setups. The cost case for 5-axis starts when the part has angled faces, deep intersecting passages, or a tolerance stack that survives only if the number of setups drops.
We run 16 simultaneous 5-axis machining centers alongside 27 three-axis machines, so the setup choice follows the geometry rather than the shop's default. A block with simple external faces does not get booked on a 5-axis machine just because one is free.
Aluminum, iron, and how each one cuts
Most blocks we see are aluminum: 6061, 6061-T6, 7075, A356 and ADC12 die castings. Aluminum cuts fast but moves with heat. A 100 mm bore can grow 0.02 mm from a 40 °C temperature rise, which is four times the tolerance on a main bore. Coolant delivery and a dwell before final measurement are not optional.
Cast iron blocks cut differently. The material is stable and damped, so roundness holds easily, but the graphite dust is abrasive and tool life drops. Ceramic inserts or coated carbide at moderate speed keep the bore size predictable across a run.
Billet blocks in 7075 or 4140 let you skip porosity concerns entirely. The trade is cycle time. Hogging a block from 150 mm plate can take hours before the first finish pass, and the material cost per block is higher than a casting. Billet makes sense for prototypes and low-volume custom work.
Inconel and titanium show up in small engine and motorsport work, not production blocks. Both work-harden, so we keep radial engagement low, use climb milling, and never let the tool rub. Cutting speed drops sharply, and the cycle reflects that.
How the numbers get verified
A block is measured on a CMM with the same datum scheme used in machining. Bore roundness gets checked at three heights, not one, because a tapered or barrel-shaped bore passes a single-plane check. Deck flatness is mapped across the full face, since a low spot in the middle of a gasket path matters more than a corner.
Surface finish is checked with a portable profilometer against the drawing callout. A deck at Ra 0.8–1.6 μm gives the gasket something to bite. Polish it to Ra 0.2 μm and the gasket can slide, which is a real failure mode on high-boost engines.
We inspect 100 percent of parts before shipment: raw material check, in-process monitoring, and final inspection. Reports go out on request. On a block, the in-process steps catch a bore that has drifted before the next operation adds more value to a part that will be scrapped.
First article inspection is where a new block program gets proven. If the first article is off, the fix is usually in the fixture, not the program. Changing the toolpath to chase a datum error hides the problem until the next order.
Casting versus billet, and when each one fits
Pick the starting form before you pick the toolpath
| Factor | Cast block | Billet block |
|---|---|---|
| Typical stock per face | 1.0–3.0 mm | 10–40 mm |
| Cycle time driver | Fixture count and datum count | Metal removal volume |
| Best for | Production runs, OEM geometry | Prototypes, custom port layout |
| Porosity risk | Possible at core shifts | None, material is wrought |
| Wall thickness | Set by the mold | Fully open to design |
| Setup count | 3–5 typical | 2–4 typical |
| Cost curve | Low per part at volume | Flat, driven by machine hours |
What each feature has to hold
| Feature | Typical target | Why it matters | Failure mode |
|---|---|---|---|
| Cylinder bore roundness | 0.005–0.010 mm | Ring seal and oil control | Blow-by, oil consumption |
| Deck flatness | 0.02–0.05 mm overall | Head gasket seal | Coolant leak, compression loss |
| Cam bore alignment | 0.015 mm over length | Bearing load and cam timing | Bearing scuff, valve timing drift |
| Main bore size | ±0.005 mm | Crank bearing clearance | Bearing knock, oil pressure loss |
| Deck surface finish | Ra 0.8–1.6 μm | Gasket bite | Seep or gasket migration |
| Bore-to-bore spacing | ±0.015 mm | Piston and rod alignment | Side loading, uneven wear |
Setup count and what it buys
| Setup approach | Holds | Best fit |
|---|---|---|
| 3-axis, 3 setups | ±0.025 mm on related faces | Simple open blocks, low volume |
| 4-axis, 2 setups | ±0.015 mm on rotated features | Blocks with side passages at 90° |
| 5-axis, 1–2 setups | ±0.005 mm on compound faces | Complex decks, angled galleries |
| Mill-turn cell | ±0.010 mm on round features | Bearing bores plus face work |
Which route fits your block
For a production block with OEM geometry, start from a casting and budget for 3–5 setups. For a prototype, a one-off, or a block with custom port and gallery layout, start from billet and accept the longer cycle. If the tolerance stack only survives under a single setup, 5-axis is the cheaper answer even at a higher hourly rate.
Common questions
How much stock should a casting leave for machining?
Between 1.0 mm and 3.0 mm per functional face is workable. Below 1.0 mm, sand inclusion or a shifted core can leave you short on one face after cleanup.
We check the raw casting before quoting. If a face is already at 0.6 mm and the core is off by 0.5 mm, the block may not clean up, and that is better to know before the program is written.
Can a block be machined in one setup?
Only if the geometry allows tool access from one direction. A straight block with open sides and no undercuts can sometimes be finished in two setups, not one.
Anything with a deck face plus a crank bore usually needs at least two. The crank bore and the deck are on opposite sides of the part, so the block has to be re-located.
What tolerance can 5-axis hold on a deck face?
We work to ±0.005 mm on related features where the setup and the material allow it. On a large aluminum deck, thermal movement can eat part of that margin, so we dwell and measure before the final pass.
Flatness targets are usually looser than size targets. A deck at 0.02–0.05 mm overall flatness seals fine with the right gasket.
Does billet cost more than a casting?
Per part, yes, once volume is high. The material is more expensive and the cycle is longer because you are removing a large volume of metal.
At low volume and one-off prototype counts, billet often wins because there is no pattern, no mold, and no minimum order. There is no minimum order quantity on our side, so a single block can run.
How do you handle a block with unknown history?
We map it first. The block gets scanned or probed to find the actual stock on each face, then the program is adjusted to the real geometry rather than the nominal drawing.
This matters on a used or repaired block where a previous deck cut already removed material. Cutting to the original drawing height would go through the deck.
What surface finish does a cylinder bore need?
A plateau hone is a separate operation from CNC boring. We machine the bore to size and leave the hone allowance, then the bore is honed to the ring manufacturer's spec.
The CNC step controls roundness and straightness. The hone controls the crosshatch and the final Ra. Both numbers end up on the inspection report.
Send us your block drawing
Upload the model or drawing and we return a quotation with a free DFM analysis within 12 hours. Production can start within 24 hours, and parts ship in 3–5 days.
12-hour quote100% inspectionNDA on request
More from the shop floor
We publish setup notes, tooling trials and inspection data from the factory floor.
