CNC Rotary Transfer Processing Guide

• 1
  Key takeaways
• 2
  How a rotary transfer cell actually cuts metal
• 3
  Cycle time math engineers should run before quoting
• 4
  Part features that fit rotary transfer and features that do not
• 5
  Rotary transfer vs 5-axis machining: which process wins
• 6
  Why repeatability holds on a transfer cell
• 7
  Tooling, changeover and what to plan for
• 8
  Rotary transfer vs 5-axis at a glance

A rotary transfer machine is a ring of fixed machining heads arranged around a rigid indexing table. Instead of moving one spindle to every feature on the part, the machine brings the part to a series of dedicated stations. Each station is set up to do one job: face milling, drilling, boring, tapping, reaming, or turning a diameter.

The part sits in a fixture that clamps once. The table indexes it from station to station by a fixed angle, usually 30°, 45° or 60° depending on how many stations the machine has. Because the fixture never lets go, features that need to hold a tight position relative to each other are generated from the same datum through the whole cycle.

The unit heads are not the same as a machining center spindle. They are built for a fixed stroke and a fixed depth, often driven by cams on mechanical versions or by servo ball screws on CNC versions. That limits how much you can change a cut on the fly, but it also means the machine can be fed hard and fast with very predictable repeatability.

A CNC rotary transfer machine still has full CNC control on the axis positions. What it lacks compared to a 5-axis center is the freedom to tilt the tool and reach a five-sided part in one setup without dedicated fixturing. The trade is deliberate: less flexibility, far higher output for the parts it is built for.

• 1
  Fixed headsEach unit carries one operation, so setup repeats are short once the cell is dialed in.
• 2
  Single clampThe part is never re-fixtured, which removes stack-up from multiple setups.
• 3
  Index timeTable rotation is measured separately from cut time and can run under one second on some designs.

The number that drives cost is the effective cycle time, not the sum of all operations. If a part has eight operations spread across eight stations, and the slowest station takes 9 seconds, the machine completes a part roughly every 9 seconds plus index time. Six other stations finish early and wait.

This changes how you balance a process. You do not optimize every cut equally. You find the station that takes the longest and ask whether it can be split across two stations, or whether the tool can be replaced with a higher-feed grade. A 20% cut on the bottleneck station is a 20% cut on the whole cell.

Index time matters more than people expect. On a mechanical machine with a fast cam index, table rotation can be a fraction of a second. Add that to the station time and convert to parts per hour. For a 9-second station plus 1-second index, that is 6 parts per minute, or about 360 parts per hour before any downtime.

Then subtract realistic availability. Tool changes, chip clearing, bar feed reloads and planned maintenance all pull the number down. Plan the cell at 75–85% of the theoretical rate for a first-pass capacity estimate, and let real data replace that assumption after the first production week.

• 1
  Bottleneck ruleImprove the slowest station first. Everything behind it is already waiting.
• 2
  Index overheadKeep it under 1 second if the station time is already short.
• 3
  AvailabilityUse 75–85% of theoretical output for planning, not 100%.

Rotary transfer suits parts that are roughly rotationally symmetric or compact in one plane, with many features on a few faces. A hydraulic manifold body, a fuel injector housing, a valve block, an electrical connector shell or a small gearbox housing are typical. They carry a family of holes, faces and threads that repeat on every unit.

It struggles with large, thin or geometrically awkward parts. A 4,000 mm welded frame or a thin-wall enclosure needs a gantry mill or a 5-axis center, not a transfer cell. The fixturing on a rotary table assumes the part can be clamped rigidly at a stable location and survive the index without deflecting.

Tolerance is where the process earns its place. Because the part is clamped once and every station shares the same datum, position between features can hold around ±0.005 mm on a well-maintained cell. If your drawing calls for ±0.05 mm on hole position, you are paying for accuracy you do not need.

Material matters too. Aluminium, brass, free-machining steel and stainless grades such as 303 and 17-4PH cut cleanly at high feed. Titanium and Inconel are possible but the tool life and cutting force make the economics harder. Plastic parts rarely justify a dedicated cell.

• 1
  Good fitCompact, clampable, multi-feature parts with stable walls and repeating operations.
• 2
  Poor fitLarge frames, thin walls, one-off geometry, loose-tolerance parts.
• 3
  Material ruleFree-machining metals reward the cell. Hard alloys rarely do.

A 5-axis machining center is flexible. It can tilt the tool, reach undercuts, and run a part from raw stock to finished geometry in one setup. It needs no dedicated cam plates or station-specific heads. That flexibility is exactly why it loses on throughput for a high-volume part with a stable design.

A rotary transfer cell is dedicated. The tooling is built around one part family, and once the cell is dialed in, cycle time is short and repeatability is high. The cost is front-loaded into fixture design, unit head setup and prove-out. That investment only pays back when the annual volume is high enough to spread it across thousands of parts.

A useful rule: if the part will run for less than about 2,000 pieces a year, a 5-axis or mill-turn center is usually the better commercial choice, because the transfer tooling cannot amortize. Above roughly 10,000 pieces a year with a stable drawing, the transfer cell wins on piece price and on consistency.

The two processes are not competitors so much as stages. Many programs start on a 5-axis center for prototypes and pilot builds, then move to rotary transfer once the design is frozen and demand is confirmed. That staging avoids spending on dedicated tooling before the part is final.

• 1
  Choose 5-axisLow volume, changing design, complex angles, one-off prototypes.
• 2
  Choose rotary transferHigh volume, stable design, many features on a few faces.
• 3
  Stage itPrototype on 5-axis, then transfer the frozen design to a cell.

Repeatability comes from removing variables. On a manual or 5-axis process, every re-fixture is a chance for a chip to sit under a locating pad or for a clamp to seat differently. On a transfer cell, the part is clamped once and indexed, so the datum is the same for every operation.

The unit heads are also rigid in a way that a tilting spindle is not. A fixed head with a short, stiff quill deflects less under cutting load, which means the tool holds size longer and the surface finish stays consistent across the run. That is why a cell can hold a fine finish without a separate finishing pass on many parts.

Thermal behavior is simpler too. The machine runs one cycle shape all day, so the thermal signature stabilizes early. There is no long warm-up ramp for a five-axis spindle orientation change. This helps when you are holding a tolerance like ±0.005 mm across a full shift.

Inspection still matters. We run raw material checks, in-process monitoring and a final inspection before shipment, with reports available on request. A stable process is not a substitute for measurement, especially on the first article and at each tool change interval.

• 1
  One datumNo re-clamping between operations, so stack-up stays small.
• 2
  Stiff headsShort-stroke unit heads deflect less under load.
• 3
  Stable thermal stateRepeating the same cycle keeps the machine at a steady temperature.

Tooling is the real project. Each station needs a head, a tool holder, a cutting tool and often a dedicated fixture nest. On a CNC rotary transfer machine the heads are servo-driven and the positions are programmable, which makes tweaking a depth or a feed easier than on a purely mechanical cam machine, but the physical tooling is still part-specific.

Changeover from one part to another takes planning. If the parts are in the same family and share a fixture envelope, changeover can be hours. If the fixture and heads must be rebuilt, it is closer to a new project. That is why rotary transfer is a poor fit for high-mix, low-volume work and a strong fit for a stable family.

Chip control deserves attention. With many heads cutting at once, chips are generated fast and must clear the fixture before the next index. Through-tool coolant and air blast are common. If chips pack into a blind hole, the next station may start on a contaminated surface.

When the cell is running, capacity is straightforward to plan. Our three plants and 127 high-precision CNC machines give us the flexibility to run transfer work alongside 5-axis, mill-turn and turning cells, so a program can scale without moving to a different supplier.

• 1
  Tooling firstBudget the head, holder, cutter and fixture nest for every station.
• 2
  Family changeoverSame fixture envelope means hours. New tooling means a new project.
• 3
  Chip clearingThrough-tool coolant and air blast keep the nest clean between indexes.