Faro CNC machine tool proficiency: the measurement loop behind 5-axis work

• 1
  What Faro CNC machine tool proficiency means on the shop floor
• 2
  How the measurement loop closes between arm and spindle
• 3
  Which part geometries justify arm-based verification
• 4
  Where arm-based measurement stops being the right answer
• 5
  Material and thermal effects you have to account for
• 6
  Machine capacity and how the arm fits our 5-axis cells
• 7
  Measurement method selection by part and tolerance

A Faro arm is a portable articulated coordinate measuring machine. Six or seven joints carry an encoder at each axis. The controller resolves every joint angle into an XYZ point in the arm's own coordinate frame, then transforms that point into the part frame using the alignment you built. No gantry, no granite table. It goes to the part instead of the part coming to it.

Faro CNC machine tool proficiency, in practice, is the ability to build that alignment correctly and to read the result against the CAD model without fooling yourself. A 5-axis machine can cut a compound-angle pocket that no caliper will ever check. The arm measures the surface, the software compares it to nominal, and you get a number in micrometers.

The value is the timing. A CMM sits in a temperature-controlled room and usually runs after the part leaves the machine. An arm can be walked up to a 4,000 mm titanium frame still clamped on the table, measured in the same setup, and the offset corrected before the finishing pass. That is where rework disappears.

One boundary up front. The arm measures points, not a whole surface. If your drawing calls out a profile tolerance over a long sweep, sampling density matters more than the arm's single-point accuracy. Plan the point set before you touch the part.

The loop has four links: probe, align, compare, correct. Probing takes a few points on a datum face, a bore, or a machined boss. Alignment uses those points to place the CAD model in the same space as the physical part. Comparison produces a deviation map. Correction translates the map into a work offset or a toolpath shift.

Alignment is the link that fails most often. A three-point plane fit on a rough casting is only as good as the surface. If the datum face was saw-cut and never machined, the fit will move by 0.05 mm or more depending on where you touch it. Machine the datum first, then align to it.

Comparison needs a tolerance, not just a number. A deviation of 0.03 mm on a bearing bore is scrap. The same 0.03 mm on a non-critical cosmetic radius is fine. Our standard machining tolerance is ±0.005 mm, but that figure applies to the dimensions the drawing controls. Every other surface just needs to be right enough to do its job.

Correction is the payoff. On a 5-axis part, a small angular error at the fixture can push a feature far out of position at the end of a long arm. Re-probing after the first article and shifting the work offset often recovers the whole batch without touching the program.

The loop is only useful if it runs before the part is unclamped. Once the vise opens, residual stress can move the geometry and the measurement you took no longer describes the part that ships.

Freeform surfaces are the classic case. Turbine blades, impellers, and organic implant shells have no flat face to sit on and no feature a height gauge can reach. A portable arm with a 2.5 mm or 3 mm probe tip can trace the surface directly and compare it to the NURBS model.

Deep cavities and undercuts are the second case. A five-axis tool can reach inside a housing through an angled entry. The same geometry blocks a fixed CMM stylus unless you build a custom fixture. The arm reaches in from the same direction the tool came from.

Thin-walled parts are the third case, and the hardest. A 1.5 mm aluminum wall deflects under probe contact force. Touch it with a hard probe and you measure your own stylus pressure, not the wall. Here we lower the probe force, take more points, and average, or move the check to a non-contact scan.

Large weldments and frames are the fourth case. A 4,000 mm machined frame cannot be moved to a CMM without a crane and a truck. The arm travels to the machine. That single logistical fact decides the method for most of our large work.

What does not need the arm: a simple turned shaft with five diameters, or a plate with a bolt pattern. Calipers, micrometers, and a height gauge are faster, and the operator can check them at the machine without waiting for anyone.

Accuracy has a floor. A portable articulated arm is typically specified in the tens of micrometers over its working volume, and that figure degrades as you extend the reach and as the joints warm up. For a feature controlled at ±0.005 mm, the arm is a screening tool, not the final referee. A temperature-controlled CMM or a dedicated gauge is.

Volumetric accuracy is not the same as repeatability. A single point touched twenty times may repeat within a few micrometers while the absolute position drifts because of thermal growth in the arm's carbon tubes. Let the arm soak in the shop environment before a critical check. Twenty minutes is not too long.

Sampling density is the second limit. A scan of a 300 mm freeform surface at 0.5 mm point spacing produces a lot of data and a slow report. Coarse sampling on the same surface can miss a local bump that the customer's own inspection will find. Agree the point plan with the customer before the first article.

Surface finish interacts with measurement. On a bead-blasted or anodized surface, probe contact reads the texture as well as the form. Measure before finishing when the drawing controls form, and after finishing only when the drawing controls the coated surface.

Finally, the arm does not replace a first-article inspection report. It produces data. The report still needs the drawing callouts, the datum scheme, and a signature.

Aluminum moves. A 6061-T6 part measured at 24 °C and inspected at 20 °C will differ by roughly 0.023 mm per 100 mm of length. On a 300 mm bore that is nearly 0.07 mm, which is larger than the machining tolerance. Note the temperature at measurement and, when the drawing requires it, normalize to 20 °C.

Titanium and Inconel are worse for the cutting process and better for stability. TC4 (Ti-6Al-4V) has low thermal conductivity, so heat concentrates at the cutting edge and the tool wears quickly. In-process probing catches the resulting size drift early. Inconel is similar, with more tool pressure.

Stainless 17-4PH in the H900 condition is dimensionally stable but hard on probes. Use a ruby tip and keep the contact force low. Beryllium copper and magnesium AZ31B are soft enough that a heavy touch leaves a mark on a finished surface.

Stress relief matters more than any measurement. A part machined from a plate without stress relief can move 0.1 mm after unclamping, and no amount of probing will fix geometry that changes after the check. Rough, relieve, then finish.

We machine more than 50 metals, from 6061 and 7075 to 316L, 4140, and TA2. Each one has a different thermal signature, and the measurement plan should reflect it.

GreatLight runs 127 high-precision CNC machines across three wholly-owned plants, with 7,600 m² of floor space and 150 technicians. The 5-axis group is 16 simultaneous machining centers, alongside 12 four-axis mills, 27 three-axis machines, and 16 mill-turn centers.

Work envelopes cover the range. Large travel is 4,000 × 400 × 150 mm. Medium cells run 750 × 1,150 × 550 mm and 600 × 600 × 600 mm. Compact cells run 500 × 500 × 450 mm and 500 × 310 × 200 mm, with a Ø400 mm rotary table available.

The arm is most useful on the large and medium cells, where a fixture error at one end of the table produces a displacement at the other. On the compact cells, a probe in the spindle often does the same job faster, and we use it.

Inspection is 100% before shipment, with raw material check, in-process monitoring, and final inspection. Reports are available on request. Certifications held are ISO 9001:2015, IATF 16949:2016, ISO 13485:2016, and ISO 27001:2022.

Quotation and a free DFM analysis come back within 12 hours. Production can start within 24 hours, and parts ship in 3–5 days. The historical late-delivery probability is below 2%.