“You need to find the root cause instead of fixing the symptoms,” said Jeff Watson of Custom Chemicals of Texas. “Troubleshooting methods apply to anything, whether it's manufacturing issues or an issue with the grass in your lawn. You have to look at the problem you have and define it as clearly as you can.”
In “Structured Troubleshooting: A Disciplined Approach to Finding and Resolving Paint Issues,” Watson said a common language should always be used between operators, management, and the supervisor.
“If you have a specific issue, always call it the same thing, so that one doesn't call it a fisheye and the other a crater, and you don't know which one they're talking about,” he said. “Make sure you're looking at a cause, not a symptom. If you have a certain particular configuration that always has a certain problem and always is in the same area, you pretty well know the problem is on a part itself and not a system problem on everything that's going through your washer. You can rule out a lot of causes before you have to do anything.
“A lot of times on painted parts, you have trash because of rejected parts. You'll have chunks of stuff in the paint that are defects. It's one thing to be able to find out whether trash is on top of the paint or underneath it. That's easy enough. If it's covered with paint, you painted over it. Start looking at whether the trash is airborne or something that's falling on a part, and if it's in the wash operation where you're actually putting it on the part because one of stages is too dirty.”
Watson said the process of conversion-coating parts is pretty straightforward. A series of steps involves conveying the part to, and through, the pretreatment system.
He said the washer is a very simple system: Clean the parts, coat the parts, rinse the parts, and seal the parts. Each stage does a specific job, and visual cues after each stage provide an indication of how well each stage is performing.
“For example, everyone knows the water break after the cleaning stage is bad — it means the parts are not clean,” he said. “But there are many other visual cues to washer performance that can tell you a lot about the condition of the process: inter-stage drying, drips; foam hanging on parts; uneven coatings or cleaning; and run-off from parts.
“Generally, alkaline cleaners clean organics better, and acidic cleaners work better on inorganic soils. Heat, the exact nature of the soil, buffering effects from the water, spray impingement, and other factors affect cleaning, and must be eliminated as factors when looking for causes of cleaning problems.
“Interestingly, some oils clean much better at lower temperatures than typical cleaners run. This is because the volatile portion of the oils flashes under higher heats, leaving the actual lubricant components behind. These components can be really hard to remove.”
How clean must the water be? He said rinse tanks hit an equilibrium within one shift. Soil loading, contamination, and overflow rate determine the equilibrium point. A spreadsheet is available to perform “what-if” and to assist in backing in to find soil-loading numbers.
Coating stages come in all forms, including iron, zinc, and chromes. But the primary function of all of these coatings is to provide a corrosion-resistant, relatively rough, inert base for subsequent coatings.
“The coating stage just coats the part,” he said. “The part must be clean and rinsed. Drag-in will buffer any coating product. Heat accelerates any reaction. Impingement makes a difference.
“Final rinses have several different concepts, and the explanation of their benefits varies with who's selling it. The important thing to remember is that the final rinse stage's purpose in life is to leave the part in its optimal condition for coating. Final rinse stages come in a few variations: surfactants to assist sheeting; corrosion inhibitors; adhesion promoters; and combination products.”
When a problem arises, Watson said it's important to ask:
- What is the defect or problem?
- Where is the earliest point it is evident?
- When does it occur?
- How does this type of problem happen?
Isolate the problem by answering:
- What is the earliest point this problem occurs?
- What types of things can cause this?
- What can we rule out?
- What have we already tried?
For example, with rusty parts after washing, ask: Is the rust on the parts before they go in? Is the rust always in a specific spot? When does it happen? If the line stops, where is the first rusty part?
Where is the earliest occurrence? After the washer, after the drying oven? After line stoppages? After a specific stage in the wash? Before the washer?
“Lay butcher paper under the conveyor to see what's falling on the parts,” he said. “Hang fly-strips along the conveyor to see airborne contamination. Is it all over or only on the top? Ask the operators.
“If there is streaking from the wash, ask: Is it visible after the first rinse? Visible after phosphating? Visible on the part before the wash? Do the streaks contain solids? Water break after the first rinse? Is the phosphate the right color? How old are the rinses? Look for drips and waterfalls.
“What happens if the burner won't keep up? The nozzles plug all the time? Do you really need to design an experiment and run tests to find these issues?
“Brainstorm ideas for investigation, then prioritize the list. Be sure you're only working on one thing at a time. Work on the most obvious things first, involve your vendors and operators, and build a written list of things to check. Gather data before making any changes. Inspect the entire area while brainstorming ideas and talk to the operators. Don't eliminate possibilities yet. Write down observations and data.”
Good data is critical
He said good panel testing involves running pairs of panels for comparison and setting up logical tests. Change only one parameter per set so that when you get a result, you know why. If one panel had a problem and the other did not, you would know what the variable was to change from one to the next.
“Good data is critical — do you have it?” he said. “Historical data is critical. Written results reduce ambiguity.
“If you can define your problem and analyze it, the improvement comes in taking out the causes of the symptoms — not just covering up the symptoms. Most people's initial response is to repaint something, but all they're doing is covering up symptoms. So they end up with multiple causes of multiple problems. Once you get into a complex situation like that, you end up treating symptoms and chasing things without being able to get back to the root causes.”
Define. Define the problem. Develop a language for problem definition. Define the scope of where the problem is occurring. Use a binder to catalog defect definitions. Use pictures and descriptions. Use the same terms every time. Keep track of the causes as they are found.
Measure. Determine proper measurements to take. Measurements must be controlled to have value. Use the proper tools and methods to get useful data.
Analyze. Analyze data from multiple sources. Historical data from good times. Current data on equivalent parameters. Use Pareto analysis to prioritize possibilities.
Improve. Improve the process by taking actions. Re-enforce actions by showing results. Graphic displays carry a lot of weight. New habits are easily lost, so be vigilant.
Control. Go/No-Go standards are best. Put understandable control points in place and automate the ones that make sense.
He said tools include process mapping, flow charts, check sheets, and Pareto charts.
Check the 6 Ms:
Manpower. Training, skill, ability, color vision or vision in general, diligence in following processes, general manpower, load balancing.
Materials. Quality of incoming powder (particle sizes, trash content, specific gravity), incoming washer chemicals, incoming substrate to be coated, and air to system (oil is a no-no).
Methods. How are tasks being done? Is there a better, more effective, or efficient way? Are tasks being done in the right order?
Measurement. You get what you inspect. Are measurements inspecting the correct parameters? Are measurements being done and analyzed? Are the measuring tools capable?
Machine. Is the machine designed for the task? Is the machine capable? Is the process capable?
Mother Nature. Changes in humidity or temperature. Climate changes through seasons. Water or air quality changes. Storms, air-flow, shifts in conditions. Thresholds of temperature, humidity, and dew point. Are the dock doors open all the time?
“If you build a cause-and-effect diagram, you can always get down to a limited number of things and eliminate them pretty quickly,” he said. “If you have a trash issue and know for a fact that all washer stations are clean, you can eliminate that and move on to the next thing that could be the problem. But if you don't understand thoroughly the things that can cause the issue you're looking for, you don't know where to look.”
He suggested making a list of factors, altering those factors through a series of trials, and tracking results of trials.
Watson gave an example of a trial matrix in action. The issue: a reddish-brown stain on the inside corner, at the top and bottom.
“Several observations were made in our inspection of the process — the most notable being a yellow-brown slick floating at the entrance to the paint tank,” he said. “The dark spots are floating trash from a known issue and were discounted because we knew they were inert and only caused trash, not brown staining. The operator periodically skimmed them until the real issue was corrected.”
They evaluated the staining and came up with these possibilities: rust; oil on parts; poor cleaning; and oil in paint. They built a structured approach of concurrent testing designed to eliminate one variable in each iteration. Tests were performed concurrently to save time and production. Vendors were present and participating (paint and pretreatment).
Parameters to be tested included pretreatment in general as a system (cleaning, rinsing, phosphating, and final rinse); and paint in general as a system (paint in the tank, post-rinse, and permeate from Ultrafilters).
“To eliminate the pretreatment, we ran one set of parts straight through the entire system, and added a set of lab-treated panels behind the pretreatment, just ahead of paint,” he said. “Both sets of panels were bad, indicating that the pretreatment was not the issue.
“Tests were performed on the washer stages to determine oil content, batch conditions, rinse quality, and phosphate conditions. Also, samples were taken from the phosphate and final-rinse stages and evaporated in a lab oven to determine what types of soils were present.
“Tests were performed on the raw paint, the paint in the tank, both post-rinse tanks, and on the permeate from the UF system.”
Test 1 indicated that pretreatment was not the primary issue. Test 2 indicated that individual stages in the washer were in relatively good control and were not contributing. Test 3 indicated that a brown, oily substance was present in the paint and not being removed by the UF system.
Pretreatment was not perfect, but not the source of this issue.
The paint tank was contaminated with a brown, oily material.
The UF system was not removing the contaminant.
The brown contaminant turned darker and surfaced upon dehydration of the paint during the cure process.
The paint vendor technician provided an accelerated “permeate to drain” process while maintaining bath balance.
Paint solids were increased to encapsulate/cover the issue
The problem disappeared after a few hours of sending permeate to drain lowered the contaminant concentration.
The problem resurfaced three times over the next two weeks and was corrected each time the same way.
The unknown contaminant finally worked its way out and has not resurfaced for several weeks.
He suggested troubleshooting by fully documenting causes, effects, and cures: What exactly was the real issue? How did that affect the process? What did you do to fix it? Did your process controls or process audits show the issue? Did you fix the controls/audits?
Then pick the best available solution: How much time do you have? How much money is available? What is the return? Should you do a stop-gap or just wait and do it right?
Make one change at a time and don't rush. Run structured, documented panel tests. Walk, observe, and eliminate possible causes. Make changes on the plan and observe results.
“Never ignore the post-mortem analysis,” he said. “Document what the issue was. Document the trials that provided useful information. Document the actions taken to correct the issue. For each action, provide the resources used and effectiveness. Use common language and a structure to provide organizational learning. Then add resolved issues to a binder and use it for training.”