Touch probe, camera and optical distance sensor on one CMM axis · sensor-to-sensor offsets registered with 0.34–0.82 µm scatter, a confocal sensor with a 300 µm range that the CMM must place, and per-size MPE on a commercial multi-sensor machine
A part with tight bores, small holes and soft or mirror-like surfaces may need a stylus, a camera and a non-contact distance sensor in one program. A multi-sensor head carries them together on the CMM's measuring axis. Each sensor has its own origin, so the combined result is only as good as the measured offsets between them. This page starts with how those offsets are registered and what scatter one study achieved, then looks at what a chromatic confocal sensor adds and demands, and at the per-sensor figures a commercial multi-sensor machine publishes.
Multi-sensor probe heads and sensor packages for coordinate measuring machines: tactile probes mounted with video cameras and chromatic confocal or white-light distance sensors on one axis, and the registration of each sensor to the others. Camera imaging probes and their ISO 10360-7 acceptance tests are covered on the Vision camera system for coordinate measuring machines page; bridge, gantry and Cartesian CMMs as whole machines on the Coordinate measuring machines (CMM) page; sub-micron reference CMMs on the High-accuracy reference bridge CMM page; CMM bridges on the Bridge Structure page. Portable arms, laser trackers and stand-alone scanners are outside this page.
01
Zhao and co-authors at Tianjin University's State Key Laboratory of Precision Measuring Technology and Instruments registered three sensors fixed on the z-arm of a high-precision CMM: a video camera (VC), a tactile probe (TP) and a chromatic confocal displacement sensor (CC). [1]
The authors argue that a reference sphere has drawbacks as the shared standard for high-precision registration: a tactile probe touches only half of it, which they estimate makes the uncertainty about twice that of a whole sphere. Their alternative is a cone and a cylinder on one axis, made on an ultra-precision machine, whose features each sensor can measure and which meet at one point. [1]
Measured against a sphere, the offsets from both standards were similar and all deviations conformed to the CMM's final MPE (Es = 2 × E3); the TP-to-CC offset in z came out about 1–3 µm larger with the sphere. The camera's z value was not used, because the camera serves 2D dimensions. [1]
Editorial calculation: 0.4 mm across 2,448 pixels is about 0.16 µm per pixel at the object, finer than the 0.34–0.82 µm scatter of the camera's simulated offset to the touch probe; in this set-up registration, not pixel size, limits how well camera and stylus data agree.
| Sensor on the z-arm | Published figure |
|---|---|
| CMM | MPE E3 = (2.2 + L/400) µm in 3D; E1 < (1.0 + L/600) µm in 1D [1] |
| Video camera | Field of view 0.4 mm × 0.3 mm, 2448 × 2048 pixels, 20× magnification [1] |
| Tactile probe | MPE 0.5 + L/1000 µm, resolution 0.02 µm [1] |
| Chromatic confocal sensor | Working distance 4,500 µm, measuring range 300 µm, MPE 0.1 µm [1] |
| Simulated offset VC to TP | [−20, −10, 20] mm with SD [0.34, 0.34, 0.82] µm [1] |
| Simulated offset CC to TP | [0, 17, 10] mm with SD [0.74, 0.65, 0.65] µm [1] |
Ask how the head's sensors are registered to one another (artefact, interval, residual offset scatter) and whether registration is repeated after a stylus or sensor change.
02
A chromatic confocal sensor measures distance without contact, but only inside a short range at a fixed working distance, so the CMM axes do the coarse positioning. The Tianjin sensor had a 300 µm range at 4,500 µm working distance. [1]
ZEISS offers its DotScan chromatic white-light sensor on the O-INSPECT multi-sensor machine for sensitive, reflective or low-contrast surfaces that make other sensors difficult to use. [3]
Editorial calculation: with a 0.3 mm range, the sensor's origin must be known to well within 300 µm before a scan starts, which is the job of the registration above; the 1–3 µm z difference Zhao et al. found between standards is about 0.3–1% of that range. At 30 kHz, a scan at 10 mm/s would take 3,000 readings per millimetre, one every 0.33 µm; the speed is an illustration, not a published figure.
| Micro-Epsilon confocalDT figure | Published value |
|---|---|
| Measuring ranges with the IFC2465/66 controller | 0.3, 1, 2.5, 3, 6 and 10 mm [2] |
| Linearity and measuring rate (IFC2465/66) | 0.1 µm; 30 kHz [2] |
| Resolution | Below 1 nm [2] |
| Light spot | Smaller than 3 µm, attributed to high numerical aperture [2] |
| Measuring angle (IFS sensors) | Up to ±70° [2] |
| Surfaces | Virtually surface-independent, including mirrors and glass [2] |
Give the surfaces that must not be touched, their reflectivity and slope, the height range to be scanned, and the point spacing needed.
03
ZEISS's O-INSPECT combines a Discovery.V12 zoom camera, VAST XXT tactile scanning and the optional DotScan sensor, and publishes length error by measuring volume. [3]
The tactile options are VAST XXT TL1, with radial stylus length up to 40 mm (star stylus) and tip diameters of 0.1 to 8 mm, and VAST XXT TL3, up to 65 mm and 0.3 to 8 mm. Camera options run from standard 100 (87 mm working distance, 0.9 µm sensor resolution) to scout 240 (30 mm, 0.4 µm); ZEISS states that scout 160 and scout 240 give 1.6 or 2.4 times the optical resolution of standard 100, and that pneumatic vibration insulation is essential for applications with scout 240. [3]
Editorial calculation: for a 100 mm length, the 3D MPE is 2.4 + 100/150 ≈ 3.1 µm on the smallest machine and 1.9 + 100/250 = 2.3 µm on the middle size. The 0.9 and 0.4 µm camera resolutions give a ratio of 2.25, not the 2.4 ZEISS states for scout 240; ask which applies. The camera as an imaging probe, and how it is tested, is the subject of the Vision camera system page.
| Measuring volume | MPE(E) 1D | MPE(E) 2D | MPE(E) 3D |
|---|---|---|---|
| 300 × 200 × 200 mm | 1.6 µm + L/200 | 1.9 µm + L/150 | 2.4 µm + L/150 [3] |
| 500 × 400 × 300 mm | 1.4 µm + L/250 | 1.6 µm + L/250 | 1.9 µm + L/250 [3] |
| 800 × 600 × 300 mm | 1.5 µm + L/250 | 1.8 µm + L/250 | 2.2 µm + L/250 [3] |
Ask for the MPE of each sensor separately at your feature lengths, and the stylus length needed to reach the deepest feature.
04
Paste it into your enquiry below or into an email to a maker, then replace every bracket.
Parts: [drawing; materials; surfaces not to be touched] Features: [smallest hole or edge; deepest pocket; form tolerances] Sensors: [tactile; camera; confocal or white light] Registration: [artefact; interval; residual offset] Accuracy: [MPE per sensor at your lengths] Scanning: [height range; point spacing] CMM: [make; head interface]
Manufacturer profiles associated with Multi-sensor Probe Head.
Explore company information and manufacturing capabilities, then discuss the scope of your project with a relevant supplier.
Each sensor's offset to the others is measured on a shared artefact. A Tianjin University study used a cone-and-cylinder standard and obtained simulated offset scatter of 0.34–0.82 µm between camera, touch probe and confocal sensor. [1]
Zhao et al. point out that a tactile probe touches only half of a sphere, which they estimate makes the uncertainty about twice that of a whole sphere. [1]
Micro-Epsilon confocal sensors offer spots smaller than 3 µm, measuring angles up to ±70° and measurement on mirrors and glass; the range is short, 0.3 mm on the smallest IFC2465/66 option. [2]
ZEISS gives O-INSPECT 3D length error MPE(E) from 1.9 µm + L/250 (500 × 400 × 300 mm) to 2.4 µm + L/150 (300 × 200 × 200 mm). [3]
Sources checked 2026-10-13. Sensor registration, offsets and the sphere comparison come from Zhao et al. in Sensors (2018); confocal sensor ranges, spot, angle and rate from Micro-Epsilon's confocalDT overview; MPE, stylus, camera and DotScan options from ZEISS's O-INSPECT technical data.
Ask for use case, specification boundaries, supplier type, and RFQ preparation information for this product.
Compare manufacturer profiles with relevant product and process capability.