Metal Cutting Guide: How Metal Actually Leaves the Part
This metal cutting guide explains what happens where the edge meets the workpiece — shear, heat, chip evacuation — and how those mechanics set the tolerances, finishes and tool life you can expect on a CNC. It is written for engineers and buyers who need to judge whether a part is a good fit for milling and turning.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
Key takeaways
What happens where the edge shears metal
A cutting tool does not scrape material off. It pushes a wedge into the workpiece until the metal ahead of the tip shears along a plane and slides up the rake face as a chip. On 6061-T6 aluminium that shear plane sits close to the tip and the chip comes off fast. On 316 stainless the metal work-hardens as it deforms, so the shear plane pushes ahead of the tip and the cut gets harder the longer the edge stays in contact.
Depth of cut, feed per tooth and cutting speed decide how much deformation the metal sees per pass. A light feed rubs instead of cutting. The edge cannot bite, so it burnishes the surface, work-hardens it, and the next pass is worse. This is the single most common cause of poor finish on stainless and titanium, and it is a programming problem, not a tool problem.
The shear zone is where most of the heat is generated. Roughly speaking, the softer and more ductile the material, the more of that heat leaves with the chip rather than soaking into the part. Aluminium carries heat away well. Titanium and Inconel do not, which is why they need lower surface speeds and more coolant volume rather than more pressure.
- 1Ductile metalsAluminium, brass, mild steel: long chips, heat leaves with the chip.
- 2Work-hardening metalsStainless, titanium: keep the edge engaged, never dwell.
- 3Hard or abrasiveTool steel, Inconel: carbide grade and coating matter more than speed.
Where the heat goes and how it moves the part
Cutting heat has three exits: the chip, the tool, and the workpiece. The chip is the one you want. When heat builds in the tool instead, the edge softens, wears on the flank, and starts pushing metal instead of shearing it. When heat builds in the workpiece, thin walls and long shafts move. A 0.05 mm thermal growth on a 300 mm aluminium rib is enough to blow a ±0.005 mm tolerance before the finish pass even starts.
Coolant strategy follows from that. Flood coolant removes heat but can thermal-shock carbide in interrupted cuts. High-pressure through-tool coolant does a different job: it breaks the chip and clears it from deep pockets. For deep holes and cavity work, chip evacuation usually matters more than cooling. Air blast plus a small amount of oil is often enough on aluminium.
Surface speed is the parameter most operators reach for first, and it is the one that punishes you fastest. Running 304 stainless at aluminium speeds will take the edge off a carbide insert in minutes. The fix is not a harder insert alone. Lower the surface speed, keep the feed per tooth up so the edge keeps biting, and let the chip carry the heat out.
Why a chip that stays in the cut ruins the part
A chip that is not cleared gets picked up by the next tooth and dragged across the finished surface. The result is a torn finish, a chipped edge, and sometimes a broken tool. In deep pockets and blind holes this is the dominant failure mode, well ahead of wear. Recutting also doubles the effective cutting load on the edge without any increase in material removal.
Chip form is something you can read. Long stringy chips on steel mean the feed is too light for the insert geometry. Short comma-shaped chips mean the parameters are in range. Fine powder means you are rubbing, not cutting. Blue or straw-coloured chips on steel are normal. Chips that come off black and smoking mean the surface speed is too high for the grade.
Geometry helps too. A positive rake insert cuts with less force and suits thin walls and weak setups. A negative rake insert is stronger and handles interrupted cuts in hard steel, but it pushes the part harder, so fixturing has to be stiffer. On a 4,000 mm long part, that difference shows up as chatter long before it shows up as wear.
How each metal family behaves under the edge
Aluminium 6061 and 7075 cut easily but behave differently. 6061 machines clean and takes a good finish with sharp tooling. 7075 is stronger and more abrasive, and it is less forgiving of a dull edge. Both need generous clearance angles so the chip does not weld to the rake face, a problem known as built-up edge that leaves a smeared surface.
Austenitic stainless 303 and 304 work-harden sharply. The rule is to stay in the cut: constant feed, no dwell, no spring passes at the same depth. 17-4PH in the H900 condition is harder again and often needs a coated carbide grade and a lower surface speed. Free-machining 303 with added sulphur is the easier choice when the part does not need corrosion performance beyond 304.
Titanium Ti-6Al-4V and Inconel generate heat right at the edge and hold it there. Thermal conductivity is low, so the tool sees high temperature even at modest speeds. Tool life drops quickly if the surface speed climbs. Magnesium AZ31B and AZ91D cut fast and leave a fine finish, but the fine chips are a fire risk and need dedicated handling. Copper and brass cut freely; beryllium copper needs dust controls.
When milling and turning are the wrong process
Cutting is a subtractive process, so it suits parts where you need material properties, tight tolerances and a solid cross-section. It is a poor fit for thin sheet with a lot of holes: laser or punching is faster and cheaper there. It is also a poor fit for parts with internal channels that no tool can reach, unless the part can be split and assembled.
Very hard materials push back. Above roughly 45 HRC, carbide struggles and the options narrow to ceramic or CBN tooling, EDM, or grinding. Hardened tool steel with sharp internal corners is usually an EDM job, not a milling job. If a print calls for a 0.5 mm internal radius at 55 HRC, the geometry is telling you which process to use.
Volume matters as much as geometry. One prototype and a 10,000-part run are different problems. Cutting wins on prototypes and low-to-mid volumes because there is no tooling cost. Above that, die casting or forging plus a finishing cut often beats cutting the whole shape from solid, especially when the part is large and mostly air.
Matching process and parameters to the job
Use this as a first pass. Confirm with a test cut on the actual material and setup.
| Situation | Process choice | Parameter direction |
|---|---|---|
| Thin sheet, many holes | Laser or punching | Not a milling job |
| Hardened steel above 45 HRC | EDM or grinding | Carbide will not hold |
| Prototype, complex 3D form | 5-axis milling from solid | Moderate speed, high feed per tooth |
| Deep pocket in aluminium | 3-axis with through-tool coolant | Chip evacuation first |
| Long thin shaft, ±0.005 mm | Turning with steady rest | Light radial load, positive rake |
| Stainless 304, poor finish | Same process, new parameters | Raise feed, never dwell |
| Large part, mostly air | Casting plus finish cut | Cut only the critical faces |
The short version
If the part needs tight tolerances in solid metal and no other process can reach the geometry, cut it. If it is thin, very hard, or mostly empty space at volume, pick a different process and cut only what matters.
Questions engineers ask
Why does my finish look smeared instead of cut?
Smeared or torn surfaces usually mean the edge is rubbing rather than shearing. Feed per tooth is too low, the edge has dulled, or built-up edge is welding material to the rake face.
Raise the feed per tooth, check the edge under magnification, and make sure the surface speed suits the material. On aluminium, a sharper positive-rake tool and better chip clearance often fix it without changing speed.
How do I hold ±0.005 mm on a long part?
Stiffness beats precision every time. Support the part close to the cut, keep the tool overhang short, and take a light finishing pass rather than trying to hit size in one heavy cut.
Watch temperature. If the part is warm when measured, it will move as it cools. Let it stabilise before the final check.
Does more coolant always give a better cut?
No. Coolant removes heat, but in deep pockets and holes the bigger job is clearing chips. A high-pressure jet aimed to break and flush the chip can beat a heavy flood.
On interrupted cuts in hard steel, heavy flood can thermal-shock the insert. Match the coolant strategy to the failure you are actually seeing.
Which materials are hardest to cut?
Titanium and nickel alloys such as Inconel are the hardest common materials. They hold heat at the edge and work-harden, so tool life drops fast if surface speed climbs.
Hardened tool steel above roughly 45 HRC is also difficult. At that hardness, EDM or grinding is often the more practical route.
Can you cut a part from a print with no CAD model?
Yes, a dimensioned 2D print is enough to quote and program most turned and 3-axis milled parts.
For complex 3D geometry, a STEP or IGES model saves time and removes ambiguity. We review the file and flag features that cannot be cut before the run starts.
When should I switch away from cutting?
Switch when the geometry or volume points elsewhere: thin sheet, internal channels no tool can reach, parts above 45 HRC, or large mostly-hollow parts at high volume.
Cutting wins on prototypes, tight tolerances and solid metal parts because there is no tooling cost. Above those volumes, casting or forging plus a finishing cut is usually cheaper.
Send the print, get a process answer
We review your drawing and tell you whether cutting is the right process, what parameters the material needs, and where the tolerance risk sits.
12-hour quoteFree DFM analysisNo MOQNDA on request
More from the shop floor
We publish setup notes, tooling trials and inspection data from the factory floor.
