Capability · Micro machining

Micro machining is not
small-scale machining.

Below about a millimetre of tool diameter the usual rules stop transferring. Runout becomes the process, chip load stops scaling, and the tool bends before the part does. Here is what changes, and what it means for a part you are about to have cut.

Runout stops being a tolerance and becomes the process

On a 6 mm end mill a few microns of spindle runout is noise. On a 0.2 mm cutter it is a large fraction of the tool diameter, so one flute takes the whole chip and the other takes none. The tool does not wear, it breaks. This is why micro milling is a spindle-and-holder problem long before it is a feeds-and-speeds problem.

The chip has to stay thicker than the edge is sharp

Every cutting edge has a radius, and below a certain chip thickness the tool stops cutting and starts ploughing — rubbing material aside, generating heat, work-hardening the surface it is about to cut again. Scaling a conventional chip load down with the cutter diameter walks straight into that. The feed per tooth has a floor set by the edge, not by the diameter.

The tool bends before the part does

Stiffness falls with the cube of length over diameter, so a slender cutter reaching into a deep cavity deflects away from the wall and leaves it thick at the bottom. The usual fix — a lighter cut — runs into the chip-load floor above. What actually works is shortening the reach, which is a decision about how the cavity is split, made long before anyone writes a toolpath.

Hard-milling replaces the EDM step, when the geometry allows

Cutting cores and cavities directly in fully hardened steel removes a heat-treat cycle and the distortion that comes with it, and it holds a finish that would otherwise need polishing. It does not replace EDM everywhere — a sharp internal corner smaller than the smallest practical cutter is still an electrode. Knowing which features fall on which side of that line is most of what a quote depends on.

At this scale, measurement is a real part of the cost

A feature you cannot measure repeatably is a feature you cannot hold, and contact probes deflect thin walls they are trying to measure. Optical and scanning methods have their own limits. Any conversation about micron-level work that does not also cover how the result will be verified is an incomplete one.

What we machine

Hardened tool steel
Mold cores and cavities, hard-milled to 62 HRC — the work this machine was bought for.
Tungsten carbide
Core pins and wear inserts, where a steel edge would not last the run.
Stainless steel
303, 304 and 316L. The 300 series work-hardens, so a tool that rubs instead of cutting makes the next pass harder.
Titanium
Ti-6Al-4V. It holds heat at the edge and springs back from the cutter, which is why reach and chip load matter even more.
Aluminum
6061 and 7075. Fixtures, housings, and a metal version of a part before it is committed to tooling.
PEEK, Ultem and Delrin
PEEK machining, Ultem (PEI) and Delrin (acetal): the heat stays in the cutter, not the chip, and thin walls spring back — sharp tools, light passes, and fixtures that do not squeeze.

What it runs on

CAM simulation: 5-axis milling of an aluminum part on our Jingdiao GRA100.
Jingdiao GRA100
Simultaneous 5-axisHard-mills mold cores and cavities directly in fully hardened steel.
Spindle
32,000 RPM, HSK-32Short, stiff taper — the interface that decides runout at small tool diameters.
Linear scales
0.1 µmClosed-loop position feedback rather than screw-and-encoder estimation.
Micro machining tolerance
±1 µmOn machined features. Molded parts and cavity steel are different axes with different numbers.

Got a feature you are not sure is cuttable? Show Marve the part or send it to the team, and we will tell you which side of the hard-mill / EDM line it falls on — and whether the dimension you need can be measured repeatably once it exists.