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 : October 2013
shows, it is better to be precise than accurate—accuracy is usually easily adjusted by calibration. The main reason for doing a Gauge R&R is to determine if the measurement system can adequately measure a given set of parts. In other words, can it discriminate among parts and measure significant differences? Gauge R&R also indi- cates the operator dependence of a measurement system. For example, a measurement system that uses a ruler for a gauge would most likely have high operator dependence, while a system with a digital gauge would most likely reveal lower operator dependence. Gauge R&R also aids in process control, revealing whether or not measurement data is consistent from measurement to measurement, day to day and operator to operator. All of these things together can provide confidence in the quality of data or provide the tools to implement changes that will begin yielding quality data. FLeXo Fit What does Gauge R&R have to do with flexography? In a word—everything. The list is lengthy, but some key variables include: • Plate Thickness • Plate Relief • Dot Size • Density • Ink Viscosity • Anilox Volume For the sake of this article, let’s focus on Gauge R&R ex- amples for final dot sizes on plate. To clarify the use of the word “part” throughout the remain- der of this article—the Gauge R&R method was developed by the automotive industry, so the term literally meant a part for a vehicle, such as a bolt for an enginemount or gasket. The term is still used because it is now common. In flexo, a part can be a multitude of different things. In the examples that are discussed later in this article, the parts are dots (measure size) in tone scales on masked and finished plates. Other examples of parts could be: • Plates (measure thickness or relief) • The same solid or halftone patches on print samples • Different anilox rolls • Batches of ink Procedures & cALibrAtion The first step to every Gauge R&R is defining the parts and the measurement system: What are the parts, what is the gauge and who are the operators? In order to select parts, the goal needs to be clear to ensure that what is measured achieves that goal. When selecting parts, have an end result in mind, because this determines the range of part values to be used. For example, if the goal is to determine if the gauge can differentiate over a wide range of values, then select a wide range of parts. If the goal is to differentiate among a small range of values, then select a small range of parts. Regard- less, it is critical to select parts from different time intervals and processes. Ideally, everything we make would be perfect, but that is not realistic. Flexo has what seems like an infinite number of variables, variables that come into play with platemaking and printing. Therefore a printer must choose parts that are rep- resentative of what it produces. If the measurement system is trying to catch errors, parts that are different enough for the Gauge R&R to determine if your system is capable of detect- ing those differences, must be included. The second thing that must be defined is the measurement system. What is the gauge? This will be based on the part. It could be the quality control tool routinely used to measure the selected part, or it may be a new gauge that will have its measurement capability tested. Once the part and the gauge are determined, select the operators. They should be the people who typically use the gauge in question to measure the parts selected. Select op- erators from multiple shifts to make sure the Gauge R&R can report if data is consistent across shifts. Once the parts and measurement system have been de- fined, design the experiment. The general rule is the number of parts, times the number of operators, times the number of repetitions minus 1, should be at least 30. As an equation: At least 30 data points are needed for statistical validity; more than 30 is better, but 600 would be overkill. The recom- mended number of parts is between 2 and10, the minimum number of operators is two and the minimum number of repetitions is also two. Next, enter the numbers of parts, operators and repli- cates from the previous step into statistical software such as Minitab to design the test and create a worksheet. One of the most important things to do when setting up the Gauge R&R is to make sure runs are randomized within the opera- tors (in Minitab, this option is named “Randomize runs within operators”). This means that the order of part measurements for each operator will be randomized, which is critical for the validity of the test. After the worksheet is created, it is time to collect the data, which may take several days, especially if you have opera- tors from various shifts. Before collecting any data, be sure to calibrate the gauge. Also make sure the process has been clearly defined, not only for use of the gauge, specific to the part, but also for the order, timing and data entry method the operators should follow. They must follow the random part order generated by the software and complete their first set of measurements before moving on to the replicates. If the operators are not spread out over shifts, timing may become an issue. Avoid operators going immediately after each other or completing their measurements in one sitting. There must be some time in between each operator and each replicate, otherwise you will not know how the device per- forms over time. Finally, make sure there is a clearly defined method of data entry. Give each operator a blank worksheet so he/she cannot see the data that has already been col- lected by the other operators. Once the data is collected, it can be entered into the software. When doing this, be sure that analysis of variance is selected as the analysis method. When the study results are generated, the key values can be used to analyze Gauge R&R results. 20 FLeXo OCTOBER 2013 www.flexography.org www.flexography.org OCTOBER 2013 FLeXo 21 Figure 3 Figure 2