A systematic approach to tackling paint dirt defects
CLICK IMAGE TO ENLARGE (+)
A cause-and-effect chart—also known as a fishbone chart or Ishakawa diagram—for reducing the problem of cotton fiber defects in paint.
What’s the most common paint defect in practically every paint shop in practically every plant in the world? It’s the stuff everyone calls “dirt.”
What’s dirt? It’s a cotton fiber or a paint chip or an overspray agglomeration. Usually, it’s anything that doesn’t fit into any of the other defect categories on typical inspection sheets.
It’s also something that paint shop managers and staff must work together to control.
In a nutshell, here’s our (and several others’) philosophy on how to do that:
The Paint Defect Analysis program is like a process within a process, with its own goals and objectives. The primary goal is to increase first-time quality (FTQ), direct OK rate, first run, or whatever name is given to parts that don’t have to be repainted.
This goal can be realized if the program meets the following objectives:
Establish, as a team, appropriate corrective action to eliminate (or minimize) the problems we can control.
What’s required to make such a program work?
The first thing that’s needed is an understanding of what the process and the program are. This may require education of managers and fellow employees to show them exactly what steps are involved, where the starting point should be and how long it might take to see the first result.
Also needed—and this is probably the most important requirement—is a significant commitment from plant and paint management. This means that they are willing to do all of the things necessary for the program to be a success. These include:
Giving the program time. With out a management commitment of time—time for team meetings, for process investigations, for reporting and for allowing process improvements to be realized—this program will fail.
What (or who) is a Paint Defect Analyst? He or she is the person who provides information the Paint Department Manager and the Dirt Team need to make process changes that result in process improvements.
What exactly does a Paint Defect Analyst do? Key responsibilities for this position include implementing the Paint Defect Analysis Program, collecting data on paint defects (both dirt count and dirt identification) at various points in the paint process, using statistical process control (SPC) guidelines to chart this data in a way that can be understood and used by the Paint Manager and the paint department staff, establishing and maintaining a paint defect reference library specifically for the paint department, leading (and if necessary helping to develop) a Paint Defect Reduction Team and soliciting the support of suppliers or other departments for the team and its dirt reduction activities.
That’s a lot of responsibility, and it requires a person with a specific set of key skills and abilities. These include a knowledge and understanding of the paint process, ability to communicate with all personnel in the paint department, self-motivation, computer skills (for data analysis and reporting), basic math and writing skills, basic SPC knowledge, paint defect analysis skills, basic problem-solving ability and the ability participate on a team and to present information in front of a group.
One of the first activities for a beginning analyst is figuring out where to set up data collection checkpoints. For each site, this becomes part of the permanent documentation for the program.
Establishing the number of parts to be analyzed is also important, and is based on production capability, the number of processes to be analyzed, the time available to do the analysis, and the number of parts the paint manager allows to be analyzed. In each plant the number of parts analyzed will be different.
The analyst chooses parts randomly to get a true statistical picture of the dirt problem. Some parts may have lots of dirt while others analyzed may have none; the analyst determines the average. In a plant that makes many differnt types of parts, consistent data collection becomes more difficult. The analyst must rely on the scheduling department for notification of the production changes.
The analyst must also rely on the inspection staff to quarantine parts that have been rejected for dirt, and these parts should represent all of the dirt problems for that process. This sometimes becomes a risky assumption.
Usually, the inspection staff will cooperate with the Paint Defect Analyst in identifying parts saved for analysis. Quite often it is necessary to stand in the inspection area and analyze parts as they come off the line. Skills with the shop microscope and sample preparation techniques are essential here, because, once analyzed, the parts usually have to go back on the line to be reprocessed.
Statistical Process Control (SPC) is a system for collecting data, using it to learn things about the paint process, fixing the things we think need to be fixed and then collecting some more data to make sure we fixed the right thing.
For our purpose, dirt count and identification are the most important things to measure. We also measure percent FTQ and percentage of parts rejected for dirt, but both of these depend on the inspection staff to consistently impose the quality standards set up by paint department managers or their customers (or both). Because of this, an analyst relies more heavily on dirt count and ID data for paint process monitoring.
After the analyst collects dirt counts on the required number of parts and fills in the data collection forms, he compiles this data into U-charts—line charts that plot the average number of defects per part versus time. These charts are usually plotted daily, meaning that for each day data is collected there is a point on the chart. Over time, U-charts will provide information on how big the overall dirt problem is, process stability and control, if the dirt problem is getting better or worse and whether or not a process change helped.
We can tell if a process is in control by calculating the mean (the average number of dirt defects per part for the entire time period on the U-chart) and upper and lower control limits. In a normal process, 99.73% of all of the data points should be between the upper and lower control limits. If they aren’t, then the process is said to be “out of control” or unstable.
After collecting dirt identification data, the information is compiled into a Pareto chart. A Pareto chart is a bar charts that plot each category of dirt identified versus how much was in each category. The idea is to figure out what the biggest dirt problems are. Pareto charts are also used to identify the coating layer that contains most of the defects and thus which process may contain the root causes of our problems.
Once paint defect analysis data is collected and reported, you know the extent of the overall dirt problem and what the specific dirt problems are. But you still don’t know what or where the sources of these defects are.
To find out, you can use a couple tools. The first is a Process Flow Chart, which can be as simple as a block diagram of the paint process or as complex as a blueprint of the entire paint department. The important thing is that it will lead the analyst to the part of the process that contains the root cause of a problem.
The other tool is a Cause-and-Effect Chart, also known as a Fishbone Chart or an Ishikawa diagram. This tool recognizes that there are five major components of any process—people, methods, equipment, environment and materials—and that problems in any one or more of these components can result in dirt defects.
Using a Fishbone Chart, if we ask (and answer) the question “Why?” several times, we should get to the root cause of a problem.
Q. Why do we have cotton fiber defects?
A. Because of our people.
Q. Why because of our people?
A. Because they are wearing street clothes.
Q. Why because of the street clothes?
A. Because they’re not lint-free.
Q. Why are they not lint-free?
A. Because they are 100% cotton.
Q. Why cotton?
A. Because it’s cellulose and breaks easily.
In this example, the dirt problem could be reduced by having paint shop personnel wear lint-free coveralls over their street clothes, and by teaching personnel to wear the coveralls correctly.
It’s impossible for one person—the Paint Defect Analyst—to identify and solve all of the problems in a paint shop. The analyst needs to be part of a team composed of people from all parts of the painting process. Each team member is chosen for the knowledge and expertise they bring to the group.
A typical team may consist of the Paint Defect Analyst, production hourly employees, Production Supervisor, Process Engineer, Maintenance personnel and a Quality representative. It may also include suppliers of paint materials and equipment when needed. Many plants have outsourced the cleaning of their paint shop, and these people should also be part of the Dirt Team.
Some of these people are core team members that attend every meeting; others are support members who attend only when their expertise is needed. In this way, the meetings don’t get too big and thus more difficult to control.
And, we’ve saved the most important team member for last: the paint shop manager. This person must provide a strong and visible commitment to the Dirt Team, the Paint Defect Analysis Program and all process improvement activities. As a core team member, the paint shop manager should attend as many meetings as possible and, if unable to attend, should send a delegate.
In most plants with Dirt Teams, some members may talk to other members several times a day; there’s always something going on they may need to discuss. That’s not to say that a team meeting takes place every day, although in some plants, they do.
When the program is first being implemented, it’s important for the entire team to meet at least once a week. Everyone knowing and understanding the paint defect analysis data, the status of preliminary trials or special projects and the results of process changes is a key factor to the success of the team and the program.
As the paint process becomes more stable, there will be fewer process changes to make.
The frequency of team meetings may decrease to one every two weeks. There still may be a lot going on, but it will be easier to monitor.