Makino CNC Machine Precision: How the Machine Holds Microns

• 1
  What Actually Makes Makino CNC Machine Precision Repeatable
• 2
  Thermal Stability Sets the Real Tolerance Floor
• 3
  Five-Axis Geometry: Short Tools, One Setup, Fewer Stacked Errors
• 4
  When This Class of Machine Is the Wrong Choice
• 5
  How to Verify Precision Before You Trust a Quote
• 6
  Which Machine Class Fits the Part

A machine does not hold ±0.005 mm because the brochure says so. It holds it because the loop between the cutting edge, the structure and the control closes without adding error. Three variables dominate that loop: how the frame reacts to load, how the machine handles heat, and how the control compensates for the geometry it cannot physically remove.

Makino builds around a rigid cast base and column. Cast iron damps vibration far better than a welded steel frame, so the tool does not chatter when it enters a corner at high feed. That matters most in hard materials. On 17-4PH or Ti-6Al-4V, a frame that rings will show up on the surface as Ra 3.2 μm and a short tool life.

The second variable is thermal. A spindle running at 20,000 rpm dumps heat into the housing, and the ballscrews grow as the bed warms. Machines in this class use cooled spindle housings, temperature-controlled oil, and scales that read actual position instead of counting motor rotations. Without that, the first hour of a run drifts and the last hour is out of tolerance.

The third variable is the control's model of its own error. Pitch error, straightness, squareness and backlash are mapped at build time, then corrected in the servo loop. This is why two machines with the same castings can perform differently: the compensation table is only as good as the laser interferometer work behind it.

• 1
  Cast iron dampingCuts chatter on hard materials and deep pockets.
• 2
  Thermal managementKeeps the first and last part of a run at the same size.
• 3
  Scale feedbackReads real position, not commanded position.
• 4
  Error mappingCorrects pitch, straightness and squareness in the loop.

Ask what tolerance a machine holds and the honest answer is a range, not a number. A cold machine at 7:00 am and a warm machine at 2:00 pm are different machines. Aluminum grows about 23 μm per meter per °C. A 500 mm part swinging 5 °C between morning and afternoon moves roughly 58 μm on its own, before the machine contributes anything.

This is why shops that run tight work keep the room at 20 ± 1 °C and let the machine idle through its warm-up cycle. It is also why lights-out runs are easier on a thermally stable machine: nobody opens the door, nobody changes the load, and the thermal state stays flat for hours. A stable machine can hold ±0.005 mm on a 200 mm aluminum part through a full shift.

Practical signs that thermal control is working: the spindle housing stays warm but not hot, the coolant chiller cycles at a steady rate, and the first-off inspection after a two-hour idle matches the last part from the previous run. If the first part after lunch is 20 μm off, the machine is telling you something about its thermal loop.

For buyers, this changes the question. Instead of asking for a tolerance number, ask how the shop verifies it: what temperature the room runs at, whether parts are measured after they cool, and what the inspection report includes. On a ±0.005 mm job, a report taken with a warm part on the machine is worth very little.

Five-axis motion means three linear axes plus two rotary axes, usually A and C or B and C. The benefit is not that the machine can reach strange angles. The benefit is that a short, stiff tool can reach them. A Ø6 mm end mill hanging 40 mm out of the holder deflects far less than the same tool hanging 90 mm out, and that difference shows directly in wall straightness and surface finish.

The second benefit is setup count. Every additional setup adds a datum, a clamp mark and a re-zeroing step. Machining five faces in one setup removes three or four error sources at once. On a housing with bores on two perpendicular faces, that can be the difference between a 30 μm position error and a 10 μm one.

The trade-off is that rotary axes add their own error. A trunnion that is not square to the spindle, or a rotary table with 10 arc-seconds of backlash, will throw a hole off position as the table turns. Accuracy on a five-axis machine depends on how well the rotary centers were calibrated and how often that calibration is checked. On our Ø400 mm rotary tables, we re-verify center position as part of scheduled maintenance, not only when a job goes bad.

Five-axis is also slower per cubic centimeter of metal removed than a dedicated 3-axis machine on simple geometry. If a part is prismatic, has features on one face, and needs ±0.05 mm, a 3-axis mill will make it faster and cheaper. Five-axis earns its cost when geometry is complex, when setups are many, or when tool access is tight.

A high-precision machining center is not the answer to every part. Start with the geometry. If the part is a flat plate with holes on one face, a 3-axis machine with a good vise will hit ±0.02 mm all day at a lower hourly rate. Putting it on a five-axis center adds setup time and rotary error for no gain.

Material matters too. Soft plastics and free-machining brass cut easily and do not reward a stiff frame the way titanium or Inconel do. On a POM prototype, the limiting factor is usually fixturing and thermal growth in the part itself, not the machine structure. Spending machine time on a part whose own material moves more than the tolerance is a poor trade.

Size is the other boundary. Five-axis trunnion machines have a work envelope, and a part that needs to swing through a rotary axis may not fit even when it fits on the table. Long parts often run better on a large 3-axis machine with a 4,000 mm X travel and a repositioned setup than on a small rotary machine that cannot rotate the part without hitting the column.

The honest rule: match the machine to the error budget and the geometry, not to the brand. The machine class buys you process capability. It does not buy you a good fixture, a stable room, or a correct CAD model. Those are on the shop, and they are usually where the tolerance is actually lost.

A machine spec sheet tells you what the builder claims. A verification record tells you what the shop actually does. Ask for three things. First, the room temperature the work runs at and how it is monitored. Second, the metrology used for final inspection and whether it is separate from the machine. Third, a sample inspection report showing the features, the nominal, the actual and the instrument used.

On tight work, the CMM matters more than the machine. A part held to ±0.005 mm on a machine and then measured with a caliper is not verified at all. We inspect 100% of parts before shipment, with raw material checks, in-process monitoring and a final inspection, and we issue reports on request. That is the part of the process a buyer can actually audit.

Ask how the shop handles a first article. A first-article inspection that maps every dimension gives you the error pattern, not just a pass or fail. If a bore runs 8 μm high on one side of the table, you want to know that before 500 parts are cut. Full first-article review is cheaper than a scrap run.

Finally, ask what happens when a dimension drifts. A shop that re-calibrates, re-fixtures or adjusts the process has a loop. A shop that re-cuts the part until it fits does not. The difference shows up over a production run, not on the first part.