by clicking the arrows at the side of the page, or by using the toolbar.
by clicking anywhere on the page.
by dragging the page around when zoomed in.
by clicking anywhere on the page when zoomed in.
web sites or send emails by clicking on hyperlinks.
Email this page to a friend
Search this issue
Index - jump to page or section
Archive - view past issues
FLEXO Magazine : February 2009
TECHNOLOGIES & TECHNIQUES More gloss @1 Less gloss @- FIGURE 2. The effect of changing angle of illumination. If I am supposed to make widgets that are between 23.7 and 23.8 inches long, and my yardstick could be off by up to an inch, then this is not a particularly helpful device for process control. How accurate does my yardstick need to be? The statistical pro- cess control gurus say that ideally, the measurement error of a device should not be more than 10 percent of the tolerance win- dow for it to be an acceptable measurement device. If the error of a measurement device uses up more than 30 percent of the toler- ance window, then it is considered unacceptable. Based on the tolerance windows in ISO 12647-6, Table 1 shows the requirements for a device that measures CIELAB values. RGB CAMERA ACCURACY There are a number of vendors selling web inspection systems that use an RGB color camera to report CIELAB values. One ven- dor claims an accuracy of 1.0 E, while another claims an accuracy of 0.2 E. It would appear that either system would be accurate enough to meet the ISO needs. However, there are a number of factors that limit the accuracy of colorimetric measurements that are derived from an RGB camera. For example, the design of the illumination presents TABLE 1. Accuracy requirements for a colorimeter. Color requirement Ideal Maximum allowable Color OK 0.8 E 2. E Production run 0.2SLlE 0.7SLlE a considerable engineering challenge. In order to agree with a standard color measurement device, all of the light must come in at 400 to 50 0 . Furthermore, the light must come in from at least three directions, that is, in a cone-shaped pattern. The detection of light must be made perpendicular to the surface (although the angles of lighting and camera could be reversed without chang- ing the measurement). As shown in Figure 2, if the angle of illumination is too steep, the camera will see more gloss than the standard and the mea- surement will be too bright. If the angle of illumination is too shallow, the measurement will be too dark. This so-called goniophotometric error is unfortunately not easy to correct, since the error depends a lot on the gloss of the stock. If the stock is very smooth (that is, like a mirror), then the angle must change appreciably before the camera starts to see gloss. If the stock is very matte (Le. rough) then there is also little change with angle. On the other hand, pretty much everything we print on lies between these extremes where the dependence on angle is largest. For makers of handheld spectrophotometers, getting the right angles of illumination is a challenge. For illumination of a moving web the challenge is greater. Fast web speeds and small targets mean short shutter times, so a lot of light is needed. On the other hand, fundamental principles of optics mean that keep- ing the light between 400 and 500 requires that an appreciable amount of light must be thrown away. Thus, the engineer who designs a web inspection system must balance the need for ac- curate measurement geometry against the cost of generating the appropriate lighting solution. FEBRUARY 2009 - www.flexography.org FLEXO
Sustainable Winter 2009