CNC secondary processing explained, from first cut to finished part
A milled or turned blank is rarely a finished part. This guide explains what CNC secondary processing does, which operations a given geometry actually needs, and where the limits sit. Written for engineers and buyers who have to approve a routing, not just a drawing.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
What CNC secondary processing actually means
CNC secondary processing covers every operation applied after the first setup comes off the machine. Drilling a cross-hole on a turned shaft, tapping a cover plate, grinding a bearing seat, anodizing a housing: all secondary. The part already has its main geometry. What is missing is function, fit or finish.
The word secondary does not mean optional or low value. On a hydraulic manifold, the secondary operations often decide whether the part works at all. A sealing face at Ra 3.2 μm may leak; the same face lapped to Ra 0.4 μm holds. Same casting, same milling, different outcome.
One practical rule: primary machining creates shape, secondary processing creates the interface. Interfaces are where parts meet other parts, where seals sit, where bearings spin, where current flows, where a human hand touches the surface. That is why secondary work is judged by function, not by removal volume.
- 1PrimaryRemoves most material, establishes datum surfaces and general geometry.
- 2SecondaryHits tight tolerances, adds features, changes surface or material properties.
Why a single setup cannot finish the part
A five-axis machine can reach five sides in one setup, and that removes a lot of error. It does not remove all of it. Some features sit on the opposite side of the part, some sit inside a bore too deep for the tool, and some need a process the spindle cannot perform at all.
Take a 300 mm aluminium bracket with a reamed bore on both end faces. Reaching both means either a second setup or a mill-turn cycle. The second setup adds an alignment error that has to be absorbed somewhere. Either the drawing loosens, or the process adds an in-process check.
Heat is the other reason. Roughing a 7075 block moves a lot of material and puts heat into the part. If the finishing pass runs immediately, the part is still growing. Letting it rest, then semi-finishing, then finishing, is a sequencing decision that belongs to secondary processing.
So the split is not a shop habit. It follows from geometry, from thermal behaviour, and from the tolerance the drawing asks for.
Sequencing rules that keep tolerances intact
Order matters more than most people expect. A part ground before heat treatment will move during heat treatment, and the grind is wasted. A part anodized before tapping will have coating inside the threads, and the gauge will not enter. Both mistakes are common and both are avoidable with one line on the routing.
The general sequence runs: rough machine, stress relieve, semi-finish, finish machine, then surface treatment, then final inspection. Deburring sits between finishing and coating, because a burr under a coating becomes a blister. Marking usually comes last, after the surface is stable.
Two exceptions are worth knowing. If the coating is thick, such as hardcoat anodizing, mask the threads and bores before coating rather than cutting them oversize. If the part is welded after machining, all tight tolerances go after the weld, never before.
- 1Heat treat before finishingOtherwise the finishing cut is chasing a moving part.
- 2Deburr before coatingTrapped burrs lift the coating and create pits.
- 3Mark after coatingLaser marking on a stable surface stays legible.
Surface finishing: what each process really changes
Surface finish is often specified as a single Ra number, but Ra alone does not describe a surface. Two faces at Ra 1.6 μm can behave completely differently if one is turned and one is bead blasted. Turning leaves a directional lay that can wick oil; blasting leaves a random texture that holds a coating better.
As-machined faces sit around Ra 1.6–3.2 μm on aluminium and steel. Fine milling pushes into Ra 0.8–1.6 μm. Lapping and polishing reach Ra 0.2–0.8 μm, which is the range where sealing faces and optical mounts live. Going below that on a production part usually costs more than the function needs.
Coatings change dimensions. Anodizing builds roughly half its thickness into the surface, so a 20 μm hardcoat adds about 10 μm per side. On a bore held to ±0.005 mm, that is the whole tolerance. Design the mask or the pre-coat size, not both.
Where secondary processing decides assembly
A part that measures correct on a CMM can still fail on the assembly line. Press fits need the right interference, and interference depends on the bore surface as much as the diameter. A ground bore and a reamed bore at the same nominal size behave differently under load.
Threads are another quiet failure point. Cut threads and rolled threads have different root geometry and different fatigue life. For a part that sees vibration, rolled threads are worth the extra step. For a one-off fixture, cut threads are fine and faster.
The engineering meaning is simple: secondary processing is where the drawing meets the mating part. If the mating part is not known at quote time, say so. It changes the process plan, and it is better to plan for it than to discover it at assembly.
- 1Ground vs reamedGround bores hold roundness better under press fit load.
- 2Rolled vs cut threadsRolled threads resist fatigue; cut threads are cheaper for low volume.
- 3Mating part unknownFlag it early; it changes the routing and the inspection plan.
Limits and mistakes worth knowing before you quote
Deep bores are a hard limit. A bore deeper than about four times its diameter needs a special tool or an EDM pass, and that changes both time and cost. Drawing a 2 mm hole 30 mm deep into a hardened part is a design decision that will come back.
Sharp internal corners are the second one. A milling cutter has a radius, so an internal corner cannot be sharper than the tool. If the mating part has a sharp corner, the cut needs a relief or the corner needs a radius. This is a drawing issue, not a shop issue.
Marking has a floor too. Laser marking on an anodized or blasted surface needs at least 1.5 mm character height to read reliably. Below that, contrast drops and the mark becomes a suggestion.
None of these limits are unusual. They are the normal constraints of cutting metal. The value of explaining them is that the drawing can be fixed before the first setup, not after.
Building a routing: step by step
- 1List every feature by toleranceGroup features into general (±0.05 mm) and tight (±0.005 mm) and count how many tight ones exist.
- 2Pick the datum firstChoose the surface that other features reference. Every tight feature should trace back to it.
- 3Decide the setup countIf more than two setups are needed, check whether a 5-axis or mill-turn cycle removes one.
- 4Place heat treatmentRough, stress relieve, then finish. Never finish before a thermal step.
- 5Place surface treatmentAfter deburring, before marking. Mask threads and bores that carry tolerance.
- 6Define inspectionName the features that get measured, at what stage, and whether reports are required.
Which secondary operation fits which feature
Match the feature on the drawing to the operation that can hold it. If a feature appears in two rows, the tighter row wins.
| Feature or requirement | Typical operation | Practical limit |
|---|---|---|
| Cross-hole in a turned shaft | Secondary milling on 4-axis | Position ±0.02 mm |
| Threaded holes in a housing | Tapping or thread milling | M2 and up, thread milling for M6+ |
| Bearing seat, high load | Cylindrical grinding | Roundness within 0.005 mm |
| Sealing face, low leakage | Fine milling then lapping | Ra 0.2–0.8 μm achievable |
| Corrosion resistance | Anodizing or plating | Coating thickness 5–25 μm |
| Deburred edges, no sharp corners | Tumbling or hand deburr | Edge break 0.1–0.3 mm |
| Part identification | Laser marking | Minimum character height 1.5 mm |
| Stress relief after roughing | Heat treat between passes | Dimensional shift 0.02–0.1 mm |
Cost drivers in secondary processing
Secondary operations rarely dominate unit price. They dominate the risk of rework, which is where the real cost sits.
| Driver | Low impact | High impact |
|---|---|---|
| Number of setups | One setup, five-axis | Three or more manual setups |
| Tolerance class | ±0.05 mm | ±0.005 mm or tighter |
| Surface requirement | As-machined Ra 3.2 μm | Lapped Ra 0.2 μm |
| Coating complexity | Clear anodize | Hardcoat with masking |
| Inspection level | Sample check | 100% with reports |
| Part size | Under 500 mm | Up to 4,000 mm |
The verdict on secondary processing
If the part is a prototype with general tolerances, run it in one setup and skip the extra steps. If it carries a seal, a bearing or a press fit, plan the secondary operations before the first cut, because no amount of careful milling will fix a missing grind.
Questions engineers ask about secondary processing
Does secondary processing always add lead time?
Not always. Deburring and marking can run in the same cycle as finishing. Grinding, heat treatment or plating add separate steps and usually a day or more.
When the routing is planned up front, the added steps run in parallel with other work rather than after it.
Can I specify Ra 0.2 μm on the whole part?
You can, but it is rarely worth it. Lapping a whole part costs far more than lapping one sealing face, and the rest of the surface gains nothing.
Specify the finish per face. One callout on the sealing face and as-machined everywhere else keeps cost down.
How much does anodizing change a bore diameter?
Anodizing builds about half its thickness into the surface. A 20 μm hardcoat adds roughly 10 μm per side, so a bore shrinks by about 20 μm on diameter.
If the bore holds ±0.005 mm, mask it or cut it oversize before coating. Decide which before the first setup.
When should threads be rolled instead of cut?
Rolled threads have a continuous grain flow at the root and resist fatigue better. Use them on parts that see vibration or cyclic load.
Cut threads are fine for fixtures, covers and low-load joints, and they are usually faster to program.
What inspection comes with secondary operations?
Every part gets a raw material check, in-process monitoring and a final inspection before shipment. Reports are available on request.
For tight features, name them on the drawing so the inspection plan measures the right thing at the right stage.
Can one supplier run machining and finishing together?
Yes. Keeping machining, grinding, coating and inspection under one roof removes the shipping and re-fixturing steps between vendors.
It also means one party is accountable if the plated bore no longer gauges.
Send the drawing, get a routing back
We review the drawing, flag the secondary operations it needs, and send a quotation with free DFM analysis within 12 hours.
12-hour quote100% inspectionNo minimum order quantity
More process notes from the shop floor
We publish setup notes, tooling trials and inspection data from the factory floor.
