Joint Arm Scanner for Wheel Thickness Measurement

• 1
  Key takeaways
• 2
  What a Joint Arm Scanner Actually Measures on a Wheel
• 3
  From Point Cloud to Thickness Number
• 4
  When This Method Is the Wrong Choice
• 5
  Portable Joint Arm Scanner vs Other Wheel Thickness Methods

A wheel is a thin disc with a curved face, a hub bore, and a rim profile that changes radius every few millimeters. A caliper gives you one chord at one spot. A dial gauge on a stand gives you a relative reading that depends on how well you set the stand. Neither tells you how thickness drifts around the circumference or across the spoke.

A joint arm scanner solves a different problem. The arm has six or seven rotary joints, each with an encoder, so the system always knows where the probe tip sits in space. Fit a laser scanning probe and you sweep thousands of points per second across the part. Fit a hard probe and you touch off a few hundred discrete points. The laser probe is the usual choice for wheel thickness because the surface is curved and you want coverage, not a handful of hits.

The output is a point cloud in the arm's own coordinate frame. That frame means nothing until you tie it to the part. So the first real decision in any measurement plan is not the scanner model. It is how you hold the wheel and how you define its datum.

Typical repeatability on a portable arm in this class sits in the 0.02–0.05 mm range for point-to-point work, and a bit wider for laser scans on shiny or coated surfaces. That is enough to catch a 0.1 mm thickness drift across a production run. It is not enough to certify a gauge-grade master. Know which one you are doing before you buy the arm.

Once the scan is done, the work moves into software, and this is where most measurement projects succeed or fail. The raw cloud is dense, noisy, and unaligned. You cannot measure thickness on it directly unless you first clean it and give it a coordinate system.

Digitize and align first. If the wheel has three or more reference spheres or a known fixture, register the scan to them. If not, use the hub bore and the mounting face as datums. Establishing the X-Y plane on the mounting face and the origin at the bore center gives you a frame that matches how the wheel actually sits on the vehicle.

Then reduce noise. A laser scanning probe picks up stray points from edges, reflections, and the fixture. Apply a noise reduction pass, then a filter, then uniform sampling. Sampling at 0.02–0.05 mm keeps the mesh light enough to handle while preserving the local thickness gradient.

Mesh it. Convert the cleaned cloud into a triangular mesh and save as STL. For thickness analysis you need two meshes: the upper face and the lower face, each meshed separately. Combining them into one shell before measurement usually creates self-intersections that corrupt the result.

Now measure. Thickness is the distance between the two meshes along the surface normal at each point. On a flat plate that is trivial. On a wheel with a dished face and a raised rim, the normal changes direction, so the software must project correctly or the number will read high near the rim.

Finally, define the lower surface properly. Set a flat function for the mounting face and a circle function for the central bore. With the X-Y plane as the primary coordinate system and the origin at the circle center, every thickness reading is referenced to the same geometry the machinist will use.

Export the report as a color map plus a table of minimum, maximum, and mean thickness by zone. That is what the quality engineer signs, and that is what feeds the next machining run.

Portable arm scanning is not free of trade-offs. The arm has to stand somewhere, and its working volume limits how much of the wheel you can reach in one setup. On a large wheel you may need two or three setups and a registration between them. Every registration adds uncertainty.

Surface finish matters more than most people expect. Polished aluminum and chrome-plated wheels scatter the laser. You get dropouts, and dropouts become holes in the mesh. A thin dusting of developer spray, or a matte tape on the suspect zone, usually fixes it. On a production line that is an extra step, so budget the time.

For a part that is essentially flat, a bridge CMM with a scanning head beats a portable arm on accuracy and on throughput. Use the arm when the part is large, when it cannot move to the metrology room, or when you need the data at the machine.

If your tolerance is tighter than the arm's repeatability, no amount of software will save the measurement. In that case you need a different instrument, not a better filter.

One more limit: the arm measures what exists. If the wheel is not fixtured the way it mounts in service, the scan is still a valid shape but the thickness number may not correspond to the functional condition. Fixture the part the way the vehicle holds it, or state clearly in the report that you did not.