|
In 1973, I was playing baseball for Parkside High School Eagles with Tony Dungy. That was also the year that SSPC published PA 2, its first standard for measuring paint dry film thickness.
Mind you, paint gauges had already been around for 30 years by then; SSPC, for 23.
Tony went on to star in football at the University of Minnesota, and I traveled to Ann Arbor to attend the University of Michigan. SSPC went on to make editorial changes to PA 2.
 |
|
U.S. Patent Office / backseatcuddler |
| In 1940, V. W. Breitenstein sought a patent for “an improved magnetic instrument ...to be used for the measurement of thickness of various articles, such as coatings.” Tony Dungy was still in high school when SSPC first issued PA 2. |
I graduated and took a job with the U.S. Army Corps of Engineers. Tony signed a contract to play football for the Pittsburgh Steelers. SSPC updated PA 2.
Tony won his first Super Bowl ring, and I attended my first of 22 straight SSPC conferences. Tony got another ring, and I became a member of the committee working on PA 2.
Technology Moves Down the Field
Personal computers became widely available. Digital Type II gauges were the latest and greatest. Computers and gauges could even talk to each other. PA 2 remained largely the same.
I left government service and became a consultant. Tony became the head coach of the Tampa Bay Buccaneers and then the Indianapolis Colts.
I testified in court for the first time. The Colts got better. I got better. Tony’s Colts won the Super Bowl. Tony retired from professional football. I retired from consulting.
Standard Update
And SSPC made technical revisions to PA 2—in 2012.
It was time for an update. It also appears that there were claims of so-called over-inspection and the interpretation of PA 2 requirements for determining the extent of non-conformance.
SSPC then came up with new procedures and re-issued PA 2. I helped. The document is much improved. Kudos to committee chair Bill Corbett and the members.
New Toys, Old Procedures
But what about all that cool technology we have now? The gauges are fast and easy to use. They store and analyze data. They shake hands with laptop computers and printers. And the average user is more sophisticated today, more educated, better trained, and often certified.
As I’ve tried to illustrate, a lot has happened since DFT gauges were first introduced, and PA 2 was first issued.
So how come if our technology is so much better, why do we have the same inspection procedure? Our tools are much more powerful, and yet, our procedure does not take advantage of that power.
So what would I propose? Statistical inference. Huh? Let me explain.
Operator Variables
There are several operator variables that affect the thickness of manually applied paints. For spray painting, these include standoff distance, normality of the spray gun to the surface about two axis, and traverse speed.
These operator-dependent variables, in combination with the pattern of overlap between spray passes, has a randomizing effect on paint thickness. That is, the thickness of the paint, when measured many times, should show a central tendency, with a more or less normal distribution.
The degree of variation in the paint thickness is an indication of the overall uniformity of the paint application process, and therefore is one measure of coating quality. The degree of variation in the coating thickness, when stated in terms of the standard deviation, can be used to predict (statistically infer) the degree of conformance to a thickness standard.
There would be no reason to abandon the basic framework of PA 2 in terms of numbers of 100 square foot areas to evaluate. Sample stratification, or the delineation of sub-areas and repeating structural elements, is also described adequately in the current standard.
|
|
| ABOUT THE THE BLOGGER |
Tim Race |
Tim Race began his career in paint technology with the U.S. Army Corps of Engineers, where he was nicknamed “Paint Dude” by his future wife. He later served as a consultant, working for Corrosion Control Consultants and Labs and independently. Race received the SSPC Technical Achievement Award in 1998 and was named an industry Top Thinker by JPCL in 2012. After 30 years in the industry, he is now enjoying retirement and letting his wife do the heavy lifting. |
|
SEE ALL CONTENT FROM THIS CONTRIBUTOR
|
|
Tagged categories:
Certifications and standards;
Coating / Film thickness;
Contractors;
Dry Film Thickness (DFT);
Painters;
Quality Control;
Quality control;
SSPC;
Thickness gauge
|
Comment from Brad Wilder, (10/8/2012, 8:29 AM)
|
Tim, thanks for the article. I'm glad people in the industry are thinking about how to use statistical process control in coatings QC/QA. In my relatively young career, I've seen my share of inspectors dutifully following outdated procedures when simply making use of the technology in their hands could both yield a better inspection and move the job along more efficiently. It would be a win-win for owner and contractor.
|
|
|
Comment from Tim Race, (10/9/2012, 8:55 AM)
|
Thank you Brad. I should add, that I would be happy to speak with any owners that might be interested in evaluating a new approach to measuring paint thickness.
|
|
|
Comment from Tom Schwerdt, (10/9/2012, 9:10 AM)
|
I can certainly see the time-saving advantages for relatively uniform surfaces (tanks, large repeating girders, dip tank, coil coating, et cetera) - but when you have 6 different painters painting 10-20 different surfaces in one afternoon (pretty common on an older truss bridge) - you don't really have much of any one type of area to develop a statistical model. To develop a statistically significant number of measurements for each sub-area, your statistical modeling will end up taking many more measurements than the basic PA2/100 square foot method. Painter1 tends to be light on the back of lattice bars on vertical boxes. Painter 2 is great on lattice bars, but is light on gusset plates facing South. Painter3 is heavy on flanges, but light on the upper web. Painter4 is light on the outside of that angle-iron diagonal. Painter 5 is light on the solid faces of those vertical boxes, but great on all lattice bars. Painter 6 gets everything just right - hope I can figure out which portion he painted. Heck, I'm understating the number of different surfaces on that truss - even a I-beam has 12 (or 14) different surfaces if you follow Appendix 2 in PA2. Giving a personal example: I have stood in one spot on a bridge and measured primer which ranged from ~5 to ~40 mils thick on the same diaphragm. Without moving my feet. Unusually bad? Absolutely.
|
|
|
Comment from Tim Race, (10/9/2012, 7:06 PM)
|
Thanks for your comments Tom. You seem to be quite knowledgeable about inspection. Your straw man is of course the worst case scenario. Older truss bridges such as those you describe, are in any case, very difficult to inspect. The PA2 inspection regimen does not work well for very complex structures. Some thought needs to go into the development of a standardized inspection regimen that would address such structures. Nonetheless, I believe that DFT gauges could be used as tools of statistical inference to quantify the quality of painting, even on very complex structures. The key is to deal with the complexity by unitizing it. If the truss bridge in your example has 75 unique surfaces, then there will be 75 unique units that require measurement and analysis. The beauty of statistical inference is that it allows you to make inferences, in a meaningful way, about the overall quality for each unit type, without making measurements on all of them. For example, let's say that for a given days production there will be 1000 lattice bars painted. The outside surface of lattice bars is a unique unit. Random selection and measurement on 5% of the units (50 bars) using three readings per bar (150 measurements total) could be used to represent the paint thickness on the outer surface of all 1000 lattice bars. The computed average DFT and the standard deviation could be compared against a defined quality standard to either accept or reject the unit called, outer surface of lattice bars. The major drawback that I see to this approach, is that the data for each type of unit, must be collected as a batch. This might necessitate multiple traverses over a given area to collect data on all of the unique units. That could be impractical. As you have pointed out, multiple workers adds an additional layer of complexity. Statistical inferences based on the production of more than one applicator may not be valid. In my example of 1000 lattice bars, the inspector would have to know which bars were painted by which worker. The DFT on bars painted by each worker would have to be batched separately.
|
|
|
Comment from Tom Schwerdt, (10/11/2012, 9:28 AM)
|
Tim, I appreciate the comments and additional detail. I agree that we could really use a better document than PA2 for complex structures. These structures may be a worst-case strawman for you, but they are extremely common for me. Now that I think about it, Mr. Wilder was also involved in the same two recent large, old truss repaints.
|
|
|
Comment from Brad Wilder, (10/12/2012, 8:35 AM)
|
Here is how I think about the issue. There is pretty wide agreement I think that PA-2 just isn't sufficient for complex structures. It takes some modification to make it work for a good inspection on a bridge. There are QC and QA inspectors out there all over the board in how they make it work. The best ones learn exactly what you are talking about Tom. They learn which areas are consistent problem areas and even which painters are mostly likely to have a particular type of defect in their work. Rather than the QC's looking at every square inch and the QA's coming behind and looking at every square inch again, they both learn over time the areas that are almost always within spec and the areas that need special attention.
What they are really doing at that point is statistical process control. There are certain specification limits. If the production process is within those limits with an acceptable sigma rating, the inspection process can become more about maintaining that level of control within the process rather than burning time checking everything. It might be that the process can be a whole crew or each painter might need to be considered a separate process.
Now I completely agree with you Tom that sometimes it's just not that easy. A 50-year old complex truss bridge is not a factory floor. Fun technogadgets can't replace common sense and experience. However, in a bridge project with hundreds of thousands of square feet to inspect over multiple coats, it's impossible to check everything within a reasonable about of time and expense. Fortunately there are tools out there that can help identify variability in the painting process and drive it out. All I'm advocating for is that we all as professionals look deeper into the ways to use it for the good of contractor AND owner rather than just printing out our PA-2 reports at the end of the day and calling it good.
|
|
|
Comment from trevor neale, (10/15/2012, 9:15 AM)
|
With coating technology becoming so sophisticated why not approach the DFT inspection process from an opacity viewpoint.I.E. achieve perfect hiding at the required DFT. This could ensure at least the minimum DFT is applied to all surfaces and perhaps reduce costs.
|
|
|
Comment from David Zuskin, (10/16/2012, 10:37 PM)
|
"With coating technology becoming so sophisticated why not approach the DFT inspection process from an opacity viewpoint.I.E. achieve perfect hiding at the required DFT. This could ensure at least the minimum DFT is applied to all surfaces and perhaps reduce costs."
The US Naval Research Lab is working with manufacturers to develop coatings systems with these properties.
|
|
|
Comment from Tom Schwerdt, (10/18/2012, 10:27 AM)
|
Brad - I think we're pretty much in agreement, at least at the theoretical level. David: That sounds promising for midcoats and topcoats, but I can't see it being effective for zinc primers.
|
|
|
