SPIE Startup Challenge 2015 Founding Partner - JENOPTIK Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
    Advertisers
OPIE 2017

OPIC 2017

SPIE Defense + Commercial Sensing 2017 | Register Today

2017 SPIE Optics + Photonics | Call for Papers

SPIE Journals OPEN ACCESS

SPIE PRESS

SPIE PRESS




Print PageEmail Page

Lasers & Sources

Cutting Edge Accuracy

From oemagazine April 2001
31 April 2001, SPIE Newsroom. DOI: 10.1117/2.5200104.0006

In sheet-metal-parts machining, the profit lies in the volume. Computer-numeric-controlled (CNC) laser cutters or punch presses can produce sheet-metal parts of 1.2 m X 1.2 m or more with a typical accuracy of 100 µm, but like all manufacturing equipment, they need to be monitored. The possible problems are myriad—the part could shift during punching, an operator could load the wrong tool, or the machine could simply malfunction. Close inspection of the first part manufactured in a production run can help detect these errors before faulty parts are produced.

Many sheet-metal manufacturers use conventional machine-vision systems or hand-held gauges or calipers to perform quality control (QC) inspections. Hand measuring is time consuming and introduces possible human error. Conventional machine vision systems incorporate cameras, illumination systems, scan optics, framegrabbers, and sophisticated algorithms for image analysis. The charge-coupled-device (CCD) camera technology often uses subpixel estimation and may require a clean room to protect the system from the effects of vibration, temperature changes, and contamination. Some manufacturers even dispense entirely with flat-part QC, relying on the machine operator to perform a quick visual check.

Without first-article inspection, the risk of cutting an entire flawed batch of parts becomes extremely high, and producing scrap cuts into profit. Full first-article inspection can be very time consuming, however, and the benefits of minimizing scrap can be offset by the time lost due to inspection. The challenge is to implement a QC program that does not create a bottleneck or slow down production.

silhouetting errors

Laser imaging offers a rapid, high-resolution inspection method, one that can be used by CNC machine operators right on the shop floor. Designed to monitor the shape of a flat part, this type of system consists of a visible diode laser, a detector, and a retroreflecting surface on which the part under test sits. When the laser beam scans over the part, the part surface generates primarily diffuse reflection with a small specular component, neither of which generate appreciable signal at the detector. When the beam reaches the retroreflecting surface, however, it generates a strong specular reflection that returns to the detector (see figure). Thus, the system can repeatedly trace over the boundary of a part at successive positions, generating a silhouette of the part and checking it for size and shape. Typical systems can measure a few hundred locations along the part edge each second.

A laser imaging system registers the silhouette of a sheet-metal part by registering a strong signal when the scanned beam passes over the part edge to reach the retroreflecting surface.

Though compact and economical, diode lasers generate divergent, elliptical beams that could introduce measurement errors into this type of system. By using an internal optic to circularize the diode laser output and an aspheric lens to collimate the outgoing beam, it is possible to produce a submillimeter diffraction-limited spot size, for example, 0.38 mm at a typical inspection standoff distance of 1.5 m. The laser power must remain at eye-safe levels, typically 0.5 mW.

In our system, a pair of galvanometers in conjunction with a high-speed servo system scans the beam in any arbitrary two-dimensional pattern to image irregularly shaped parts at a rate of hundreds or thousands of points per second. A scanned beam typically is configured to cover a 60° field-of-view.

A secondary adjustable focusing lens and circular wedge prism collect returned light from retroreflective background material, and a photodiode detector provides wide-band response. On a shop floor, such a system is vulnerable to contamination from ambient light. By making the detected beam path coaxial with the outgoing beam, it is possible to restrict the detected field-of-view. This factor combined with a laser line filter can eliminate interference due to external sources.

To avoid particulate contamination from the shop-floor environment, the retroreflective surface needs to be protected by a cover sheet transparent to the laser wavelength. The retroreflective material consists of a fine-grained pattern of 50-µm-diameter glass spheres, each of which serves as a miniature retroreflector. Because the material is below the surface of the part, the beam reaching it is defocused. The combination of fine-grained material and defocused beam eliminates any laser speckle effects and provides immunity to any defects in the retroreflective material.

real-world problems

This type of system cannot detect surface relief or other markings on the part. Using techniques similar to photogrammetry, the design can derive three-dimensional (3-D) information through analysis of the scanning geometry and part outline. Current systems based on this design have achieved resolutions of better than 25.4 µm at a distance of 1.5 m. Methods analogous to subpixel estimation techniques can enable the design to achieve a scanning repeatability on the order of 2.5 to 7.5 µm.

Although siting a high-speed inspection system on the shop floor next to the CNC equipment decreases inspection time, shop environments in general present problems for precision metrology equipment. Self-calibration can compensate for changes resulting from reasonable vibration or shifts in temperature, using a set of precisely located targets placed between the retroreflecting surface and the glass cover sheet as references. By viewing the target positions and calculating the 3-D position and orientation of the projector relative to the reference frame, the system can calibrate itself.

With six to 12 targets, such a system also can correct for mechanical and electronic drift in the projector. The 3-D position of the point of projection is known. If the thickness of the material is also known, as in a calibration piece, it is possible to calculate whether the beam is striking the top or bottom surface of the part and compensate for the relative position of the scanned beam.

Because the system is an optical inspection system, it will record any defect in the part shape even if caused only by transient factors such as dust or other debris on the part or supporting glass surface. (In testing, we once observed a transient part defect, which we finally attributed to the meandering of a spider mite across the part surface.) In addition, the system is quite sensitive to the edge condition of the part. Unlike coordinate measuring machines that typically measure features at the midsection of the material, this type of design will image burrs and other relatively cosmetic features of the part.

In addition, since the scan source originates from a single point, the part silhouette is formed by a combination of the top- and bottom-part edges. This can be problematic given that large dimensional differences can exist between the top and bottom surfaces of the part. Software can compensate for such problems, however, provided that information about the part manufacturing process is known. (For example, given a specification of die clearance and part material, a system can estimate the 3-D profile of the part edge.)

Laser imaging systems offer manufacturers a fast, reliable technique for volume inspection. The technology is robust enough to operate on a production floor without compromising performance or slowing down production. Less idle time means higher production and higher profit, making the investment in the system well worthwhile. oe


Kurt Rueb
Kurt Rueb is chief scientist at Virtek Vision, Waterloo, Ontario, Canada.