What Is the Function of the Machine Tool Probe?

• 1
  How a machine tool probe finds a surface
• 2
  The five functions a machine tool probe performs
• 3
  When probing helps and when it does not
• 4
  Probing accuracy versus machining accuracy
• 5
  Probing method compared

A machine tool probe is a measuring head held in the spindle or turret. A ruby or silicon nitride stylus touches the surface. The contact breaks a kinematic seat or bends a strain gauge, and the head sends a trigger signal to the control. The control freezes the axis position at that instant and stores it as a coordinate.

The machine itself becomes the measuring device. No part leaves the fixture, so there is no re-clamping error between measurement and cut. That single fact explains most of the value of probing on a 3-axis mill, a 5-axis machining center, or a mill-turn cell.

Typical touch feed rates sit between 50 and 300 mm/min for the approach move, then drop to 5–20 mm/min for the final touch. A slower touch reduces the bending of the stylus and the pre-travel error. Fast approach, slow contact: that pattern matters more than the probe brand.

Repeatability of a good kinematic touch probe lands near 1 μm. That is not the same as accuracy. Accuracy includes stylus bending, lobing, and the calibration of the probe ball diameter, which is why every probe gets calibrated against a known ring gauge or datum sphere before a job runs.

• 1
  Kinematic seatThree contact pairs; the head reseats to the same position after every touch.
• 2
  Strain gaugeNo moving seat; used for scanning and continuous measurement.
• 3
  Optical or inductiveSignal transmission for heads that spin with the spindle.

The first function is workpiece setup. The probe touches the stock on X, Y, and Z, then the control writes the work offset automatically. On a casting or forging with 1–3 mm of stock variation, this replaces a manual edge finder and a dial indicator, and it removes the operator judgment from the setup.

The second is in-process verification. Between roughing and finishing, the probe checks a boss diameter or a pocket wall. If the feature sits 0.03 mm from nominal, the control can shift the finishing offset or flag the part before more time is spent on it.

The third is tool setting. A spindle-mounted tool probe or a table-mounted tool setter measures tool length and diameter, and detects a broken edge. This matters on lights-out runs where nobody watches the first part. A broken 6 mm end mill caught in 10 seconds saves the blank behind it.

The fourth is adaptive control. The probe feeds measured stock condition back to the program, which adjusts depth of cut or feed. This is common on near-net-shape parts where the as-cast surface wanders. The fifth is final part verification inside the machine, which produces a report instead of a paper log.

• 1
  SetupFind the work offset on rough stock within seconds.
• 2
  In-process checkCatch drift before the finish pass, not after.
• 3
  Tool settingLength, diameter, and breakage detection.
• 4
  Final verificationReported dimensions from the same fixture.

Probing pays off when setup time dominates the cycle. A one-off bracket on a 3-axis machine may take 40 minutes to indicate by hand. A probe cuts that to a few minutes and frees the operator to run a second machine. On a 10,000-part run, the setup saving is amortized to nothing, so the case has to come from scrap prevention instead.

Probing also pays off on tight features. If a bore must hold ±0.005 mm and the process drifts with tool wear, an in-process check plus an offset update keeps the run inside tolerance. Without it, the operator measures offline, walks back, and edits the offset by hand.

It helps less on thin flexible walls. The touch force of 0.5–1 N bends a 0.5 mm aluminum wall, and the probe reads the deflected surface rather than the free-state surface. For those parts, an offline CMM with a low-force head is the better instrument.

It also struggles below about 0.5 mm feature size. The stylus ball, commonly Ø2 mm or Ø4 mm, cannot enter a narrow slot or a small fillet. Deep bores with a length-to-diameter ratio above 10:1 are another limit, because the stylus shank rubs the wall before the ball reaches the bottom.

• 1
  Good fitRough stock, tight bores, unattended runs, hard-to-reach datum faces.
• 2
  Poor fitFlexible walls, sub-millimeter slots, 10:1 deep bores, free-form surfaces.

A probe and a machine tool do not share the same error budget. The machine positions to its own volumetric accuracy. The probe adds pre-travel, stylus bending, and thermal drift on top. A machine rated at ±0.005 mm does not become a ±0.005 mm CMM because a probe is fitted to it.

Thermal drift is the largest practical error on a long run. A spindle warming by 2 °C over four hours moves the tool point by tens of microns. Probing at the start of the shift and again after two hours shows the drift, and the control can compensate. Without that check, the last parts of the shift wander.

Stylus length matters too. A 50 mm stylus bends more than a 20 mm one, and the bending scales roughly with the cube of length. Keep the stylus as short as the feature allows. If a long stylus is unavoidable, calibrate with it mounted, not with a short one.

For reference, GreatLight machines to ±0.005 mm (±0.0002 in) and finishes to Ra 0.2–0.8 μm where the drawing calls for it. Probing supports that work. It does not replace the final inspection, which runs on 100% of parts before shipment.

• 1
  Pre-travelThe small overrun before the signal fires; removed by calibration.
• 2
  Thermal driftGrows through the shift; re-probe a datum to track it.
• 3
  Stylus bendingScales with the cube of length; keep it short.