Machine vision inspection for the protrusion rate of a diamond tool

The performance of a diamond tool is based on the protrusion rate. The protrusion rate inspection is a costly, labor-intensive activity in industry. Normally, about 20 to 35% of the protrusion rate is provided for the final inspection of diamond tools. There are five known methods used for the protrusion measurement: mechanical dial gage, electrical dial gage, surface roughness measuring machine, focusing by microscope, and manual comparison. In this study, a machine vision system was used as an inspection tool to determine the protrusion rate of a diamond tool. The method developed is a noncontact method without manual judgment. Three sets of field samples were used to demonstrate the proposed method for determining protrusion rate.
The diamond tool is a widely used tool for hard materials, such as concrete, asphalt, all sorts of stones and glasses, optical glasses, hardened alloys, tungsten carbide, and so on. The diamond used to make the tools could be from natural diamonds, artificial diamonds, or other types of superabrasive material, such as cubic boron nitride. The application of diamond tools includes lapping, polishing, grinding, sawing, drilling, cutting, turning, and so forth. Among all applications, about 20% of the diamond tools are used for hard material (such as stones) drilling. These types of tools consist of about 50 to 60% of the revenue of the diamond tool industry (Chen 1993).

The process for making a diamond tool is first mixing the diamond grinding material with a metallic or resin-type abrasive material. The tool is then formed through a "sintering" process (see Figure 1). At this point, the diamonds have not protruded from the tool surface. The next step in the process is "truing," that is to use a SiC or A1203 sand wheel to balance the centers of the inner and outer cylindrical surfaces. This is followed by "dressing," which causes the diamonds to protrude through the surface. Chen (1995) reviewed various truing and dressing methods. The finished tools are shown in Figure 2. The surface of the tool is shown in Figure 3.

The protrusion rate (PR) is defined as the ratio of the height (h) of the diamond above the tool surface and the diameter (d) of the tool. The ideal PR is h = 1/3d, shown in Figure 4. If excessive dressing occurs, that is, h>> 113, then the diamonds fall out easily. On the other hand, if insufficient dressing occurs, that is, h

Traditionally, there are five known methods used for the measurement of PR (Busch 1989, Chen 1993, Chen 1995, Sheiko et al. 1993, Weck 1984, Lo 1992):

(1) A mechanical dial gage to rotate the diamond tool with a fixed centering device and measure the variations in the surface smoothness using a mechanical dial gage.

(2) An electrical dial gage. This is similar to using a mechanical dial gage except that the gage is an electrical gage.

(3) A surface roughness measuring machine. This device is usually used to measure in flat surface roughness, and it can be set up for measuring the protrusions in a diamond tool.

(4) An optical microscope (Brinksmeier, Hoper, Riemer 1996). A sample object (i.e., a piece of coated diamond tool) is placed under a microscope, with the focus on the high and low spots to determine the PR.

(5) Manual comparison. This is similar to the surface roughness specimens normally used in a machine shop. The PR is determined by comparing the specimens and the diamond tool surface with the naked eye.

The first three methods must use a contact-type stylus, whose size can affect the measurement results. All five methods require extensive manual work and are difficult to automate. The objective of this study was to use a CCD-- based machine vision system to determine if PR can be automatically determined with a noncontact device.

Methods and Results

The original thought was to use a vision system to measure the protrusion rate of a diamond tool and compare the results using a traditional surface profilometer. Figure 5 shows the flowchart for this process.

The vision system used in this study consisted of a CCD camera (Toshiba IK-530S) with 256 x 256 pixel resolution, a TV monitor (SAMPO BMC-1202T), and an IBM-compatible PC with a frame grabber (HRT512-8 from Caten Systems). A Misutoyo SURF-400 profilometer was used to measure the surface roughness of the diamond tool. The idea behind this was that the protrusion rate, surface roughness, and diamond area from the CCD image might have a good correlation. Figure 6 shows the hardware and software setup.