Many industries today would like to be able to measure and control the color of stained wood surfaces. Companies that make cabinetry, furniture, building products, pool tables and much more increasingly want to provide better quality control measures around the production of their products and provide a more uniform and consistent color to their customers. They know that those customers increasingly perceive consistent appearance as a measure of quality.
Everyone knows that there are natural variations within wood types regarding appearance. It’s easy to discern the tight, uniform wood grain pattern of a birch or a beech, and the significant variation within the pattern of an oak or an ash. How do we deal with this variation when we’re trying to use a spectrophotometer and establish a good color quality control program with these materials?
I recently did a quick test using a spectrophotometer on three different wood panels of various colors to determine if I could establish a repeatable test method. Here are the three panels:
Starting with the boldly grained Walnut sample on the left, single points measured randomly as a standard and a sample might show a delta E of 7 or 8 or even higher! Everyone would agree that a delta that high can not be used effectively in a color qc system. However, when I was careful to take my measurements by mostly covering the background color with the eye of the spectrophotomer, while trying to exclude the prominent graining, and I increased the number of measurements to show an average of nine, these are the results:
While a .42 delta E might not be an acceptable amount of variation with an opaque, homogenous, paint or plastic sample, it can certainly be a useful level of repeatability for accepting or rejecting stained wood products.
When I repeated the same process measuring the middle sample above, a Maple wood type, I achieved similar results, with a variation of .77 delta E. Again, this is probably too much variation in many traditional color quality control programs for evenly colored products, but if the intent is to capture most of the visually acceptable samples and reject most of those that appear to be off color, it’s a perfectly usable tolerance for quality control.
The last panel, Gunstock Oak, appeared daunting at first. With such obvious variation between the lighter and darker portions of the wood, how could I possibly get good, useful results? By sticking to the script, taking a large number of measurements (9) and doing my best to measure one shade of color (lighter, and between the grain) I achieved the following results:
Even on a sample that has obvious and significant visual differences throughout the pattern, it is possible to achieve a very useable tolerance for color quality control with a spectrophotometer. On the Gunstock Oak we achieved a delta E of 0.53, measuring it once and then against itself as a sample. Again, “good enough” for most color QC.
Oftentimes I run across a color measurement application that might at first seem impossible. In many cases, however, if I’m willing to adjust my ordinary measurement procedure a little and accept a slightly looser tolerance, I can effectively use a spectrophotomer to achieve very good results. Give it a try on your product and let me know how it goes!
Jim MacDonald is a Business Line Manager for Industrial Paint and Coatings Applications at BYK-Gardner USA.