Layout Settings for Mirror Milling Technology and Their Real-Time Use

• 1
  Key takeaways
• 2
  What Mirror Milling Technology Actually Demands From a Layout
• 3
  Vertical and Horizontal Layout: Which One Fits Your Part
• 4
  Slide Table, Pin Support and Drive: Where the Error Comes From
• 5
  Real-Time Application: Probing, Offsets and In-Process Control
• 6
  Materials and Features That Suit Mirror Milling
• 7
  Layout Settings Compared by Part Type

Mirror milling produces a reflective surface by controlling the geometry of the cut, not by polishing afterwards. That puts the burden on the machine layout. Every element between the spindle nose and the part, including the pin, the sliding table, the drive and the fixture, adds its own error to the finished surface. On a mirror finish those errors show up as visible marks, so layout tolerances are tighter than on a normal milling job.

The first question is not which machine to use. It is how the workpiece will sit and how the load travels through the support system. A part that hangs off the table edge or rests on three points will deflect under cutting force, and no amount of spindle accuracy recovers that. We check the support stiffness before we discuss spindle speed.

The second question is the surface area you actually need to be mirror-like. A full reflective face on a 600 × 600 mm plate is a different job from a Ø80 mm optical window seat. Restricting the mirror zone to the functional area cuts cycle time and reduces the risk of a rejected part.

Third, decide whether the surface is cosmetic or functional. A visible cover plate accepts Ra 0.8–1.6 μm. A light path or a seal face usually needs Ra 0.2–0.8 μm. The layout, the tool and the inspection method all change with that number.

• 1
  Support stiffness comes firstCheck deflection under cutting load before choosing spindle speed.
• 2
  Limit the mirror zoneOnly the functional face needs the fine finish.
• 3
  Cosmetic vs functionalRa 0.8–1.6 μm for appearance, Ra 0.2–0.8 μm for optics and seals.

Mirror milling machines are built in two main arrangements. In the vertical layout the pin sits on a vertical slide and the part rests on a horizontal table. In the horizontal layout the part is held on a vertical face and the pin approaches from the side. Both cut mirror surfaces. They fail in different ways.

The vertical layout is easier to load. The operator places the part flat, tightening force is low, and clamping distortion stays small. That makes it the default for small and medium flat work, roughly up to the 500 × 500 × 450 mm envelope on our compact machining centers. The limit is chip and coolant evacuation: on a deep pocket or a tall wall, debris collects at the bottom of the cut and drags across the finished face.

The horizontal layout handles longer and heavier parts because gravity pulls the part against the locating face instead of away from it. It also drains chips better. The trade-off is that the weight of the part now loads the pin and the slide in a direction they were not designed for, so the support system has to be re-verified for each new part weight.

A practical rule: if the part is under about 20 kg and mostly flat, the vertical arrangement is simpler and cheaper to set up. If it is long, heavy, or has deep features that trap chips, go horizontal and budget extra time for support verification.

• 1
  Vertical: easier loadingLow tightening force, small clamping distortion, good for flat parts.
• 2
  Horizontal: better chip evacuationSuits long, heavy parts with deep pockets.
• 3
  Re-verify supports after a weight changeThe pin and slide see a different load direction.

The sliding table sets the geometric accuracy of the whole process. If it pitches or yaws by a few micrometres over its travel, the mirror surface inherits that error as a waviness that polishing cannot remove. Most mirror machines use static gas pressure support for the pin and, increasingly, for the table as well. Liquid static pressure support is the newer direction and gives higher stiffness, at the cost of a more complex fluid circuit.

The drive system decides how smoothly the table moves. Ball screw and nut pairs are rigid and predictable, but their positioning stability depends on the precision of the screw and nut together. Steel belt and linear motor drives remove some of that mechanical stack-up and run more smoothly at low feed, which matters when the finishing pass is only a few micrometres deep.

In practice we match the drive to the surface, not to the brochure. A large flat face at Ra 0.8–1.6 μm tolerates a well-maintained ball screw layout. A small optical seat at Ra 0.2–0.8 μm benefits from a smoother drive and a lighter finishing pass.

One point that gets overlooked: thermal drift. A slide table that runs for two hours warms up and moves. On a mirror job that shows up as a gradual change in depth of cut across the plate. Warm-up cycles and in-process checks handle it better than any single machine specification.

• 1
  Table geometry sets surface wavinessPitch and yaw over travel transfer straight to the finish.
• 2
  Match drive to surface sizeBall screw for large faces, smoother drive for small optical seats.
• 3
  Warm up before the finishing passThermal drift changes depth of cut over a long run.

Real-time application in mirror milling means the machine measures and corrects while the part is still on the table. An in-process probe touches the surface, compares the result with the nominal model, and sends an offset to the control. The next pass removes the difference. This is how a mirror zone stays inside ±0.005 mm across a long plate without re-fixturing.

The correction only works if the fixture is stiff. If the part moves between the probe touch and the cutting pass, the control corrects toward a position the part is no longer in. Probe and cut have to happen under the same clamping condition. That usually means probing on the machine, not in a separate inspection room.

We pair real-time offsets with a full inspection routine: raw material check before cutting, monitoring during the run, and final inspection before shipment. Reports are available on request. The measured data tells the operator whether the layout is holding or drifting.

Real-time control also sets a practical limit. Each probe cycle costs cycle time. On a part with one small mirror seat, two or three probe points are enough. On a 4,000 mm long face, the probe grid has to be dense enough to catch waviness, and the cycle time grows accordingly.

• 1
  Probe and cut under the same clampingOtherwise the offset corrects toward a stale position.
• 2
  Routine inspection backs up the offsetsMaterial check, in-process monitoring, final inspection.
• 3
  Probe density drives cycle timeSparse points on small seats, dense grid on long faces.

Mirror milling works best on materials that cut cleanly and hold a sharp edge. Aluminium alloys such as 6061, 7075 and 6082 are the easiest starting point. They take a fine finish without loading the tool, and the layout can be simple. Brass grades like C36000 behave similarly.

Stainless steels are workable but less forgiving. Grades 303, 304, 316 and 17-4PH tend to smear rather than cut at low feed, and a smeared surface looks dull under light even when the geometry is correct. The layout compensates with a sharper tool, a lighter finishing pass, and better chip evacuation. Titanium TC4 (Ti-6Al-4V) and Inconel follow the same logic with tighter thermal control.

Plastics are a separate case. PMMA and PC scratch easily, and a mirror finish on them depends more on tool edge condition and dust control than on table stiffness. If the part is a window or a light guide, we often recommend machining to Ra 0.8–1.6 μm and leaving the final optical polish to a specialist, because the milling layout cannot deliver an optical-grade surface on every polymer.

What does not suit the process: deep narrow slots, thin unsupported walls below about 1 mm, and parts with hard inclusions that interrupt the cut. In those cases the layout cannot hold the finish, and a different process is the honest answer.

• 1
  Aluminium and brass are the easy case6061, 7075, 6082 and C36000 take a fine finish with a simple layout.
• 2
  Stainless and titanium need tighter controlSmearing and heat change the surface; lighter passes help.
• 3
  Thin walls and hard inclusions are a noBelow about 1 mm unsupported, the finish will not hold.