
By Ron Lynch, president/owner, R J Lynch & Associates
After almost 50 years involved with web-handling and consulting, I have seen a few mistakes. I must confess, for the first 15 years, I was an untrained hack and made many of them myself. Here are some of the mistakes I have seen and made as they relate to cambered webs, roller surfaces and troubleshooting, along with suggested solutions.
Editor’s Note: This technical article won the Web-Handling Committee Best Paper Award at the AIMCAL R2R USA Virtual Conference that was held online in October 2020.
Camber

Don’t neglect camber when designing new equipment. Cambered web is curved web and is longer on one edge than the other. When straightened under tension in a web path, the short edge has higher stress and strain, and the long edge has lower stress and strain. If the average tension is less than the “critical” tension, the long edge can be completely loose (see Figure 1).
Twice, I have seen disastrous consequences when camber was overlooked. In one case, a wide cambered web was slit into several narrow slits. The process demanded a low average tension and, after slitting, the long edge had no tension and fell to the floor. The company designed and built an expensive intermediate tensioning device that could only delay by a few minutes the inevitable fall to the floor.
The other case was a newly built laminating machine for a new and unusual product. The machine was well engineered and well built in all ways. The problem was there was no tension sweet spot that would prevent the laminate from curling up on one edge and down on the other edge. The strain profile across the cambered web was too much to achieve a lay-flat laminate across the full width. Only one point in the middle of the laminate could be satisfied. The immediate solution was to run at one third of the design width. Long term, the company hoped to work with the supplier to try to make a straighter web.

The radius of curvature can be measured in a cambered web, and the stress-strain difference across a fully tensioned web can be calculated using the radius and width. The critical tension can be calculated from the strain difference (see Figure 2).
The shorter the radius, the tighter the curve and the greater the strain difference across the web. Also, for a given radius, the strain difference across the web increases as the width increases. This explains why a narrower width was a solution for the second case described above.
Generally, web camber is the result of a non-level thickness profile across the width of a web. The thicker web lanes build up to larger diameters in the parent roll. The long edge of a cambered web forms when it is stretched around the larger diameter/longer circumference of the parent roll. Parent roll-diameter profile can be used to estimate camber (see Figure 3).

When designing new equipment or evaluating new materials, it is important to understand how camber, width, thickness profiles and parent roll-diameter profiles can determine the success or failure of a project.
Roller surfaces
Attention to rollers is important to achieve a reliable process. Webs like perfect rollers. Worn roller surfaces often are overlooked as the cause of problems, as is the question of why they wore and what to do to fix them. In one case, sandpaper was adhered to cores to enhance grip during splice/transfer. Over time, the sandpaper wore grooves in rollers that contacted the core, and the grooves caused wrinkles in the last few hundred feet of web on the full parent roll. The company didn’t recognize the grooves as the cause of the wrinkles. The simple solution in this case was to replace the worn rollers and remove the sandpaper.
Contamination on rollers can cause similar problems. In an acceptance test, a machine was required to make good product for one hour without making adjustments. After starting, product quality would slowly drift out of spec, and the run would be aborted. Engineers would test every component and operators would adjust and clean everything and still have the same failure on the next attempt. Eventually, we identified that coating buildup on rollers was changing their diameter and surface speed. This was enough to upset the process. The immediate and temporary solution was to not clean the machine’s rollers. Preventing the build-up was the long-term solution.
Test design
Changes without a plan and documentation are not a test. Experiments are a way to collect information to prove or disprove a hypothesis, to generate data to calibrate a model or to gather decision-making information about something. The common theme to these is a hypothesis, model or question, and a carefully designed experiment to gather specific data to achieve the defined objective.
After listening to our technicians complain about adhesive-pattern variation, and to the equipment manufacturer argue our technicians weren’t installing the parts correctly, I designed and ran a test in the manufacturer’s shop with its technicians on its pilot line that convinced the manufacturer it was the manufacturer’s part-to-part variation causing the problem, and nothing else. If you want to convince people that you are right and to do it your way, you need to have a convincing argument and evidence to back it up. Well-designed tests and data collection are how to do that. You also need to be consistently credible and not mix personal belief with supported facts.
Troubleshooting
When people don’t have an understanding of how something is supposed to work, it is hard to troubleshoot or even know there is a problem. In web handling, and just about everything else, troubleshooting is based on how things do look or work compared to how they should look or work, and how things do measure compared to how they should measure. If you don’t have the “how they should look or measure” knowledge, you are at a loss to know what to measure, fix or adjust – or that things are even wrong. I sat in a test review and one of the conclusions was that the dancer did not control tension. There was not awareness by the test engineer that there was a problem with this statement. Dancers are well-understood web-handling components, but the person didn’t have the “should work like” knowledge. The test engineer ran a test on faulty equipment, didn’t know how to adjust the equipment or didn’t know how to measure tension.
In contrast, another engineer, after much calculation, set up driven-roller speeds for a test. All zones except one delivered the expected tension. The operator simply adjusted the speed in the control panel to get the required tension. The engineer later investigated, found and fixed a programming error in the control program. The engineer had the “should look like” knowledge and followed up when it “didn’t look like” it should.
Don’t rely solely on control panels and data-acquisition systems to tell the whole story. Besides the obvious knobs on the control panels, machines can have hundreds of neglected, forgotten measurements and adjustments that determine the quality of the product and reliability of the machine. (Yes, we had knobs on control panels 50 years ago. The knobs are gone, but the hundreds of other adjustments still are there.)
The red one runs better than the blue one
As a young engineer, I remember that my decision on which technology to support likely was based on my allegiance to my manager or the opinion of a respected co-worker, not on any personal knowledge of the technology. The red one vs. the blue one referred to two rollers – one with a red cover and one with a blue.
I didn’t know much about rollers at the time but knew that the color itself probably didn’t matter. It was enough that my boss told me it ran better. I didn’t know why it was better and didn’t think to find out. It was just red vs. blue. Judging parts, processes or materials by the wrong properties or results is a common mistake. To successfully design or scale-up a process, you need to know what design properties, setup adjustments and process variables are important to the process and product. Guessing wrong can lead to problems.
Grooved rollers
Not all grooved rollers are wrinkle-removal rollers. Grooves in rollers can have two functions. In rollers with flexible elastomeric covers, the flats between properly cut and spaced grooves will deflect in an outward direction under pressure from the web. The outward deflection creates a wrinkle-removing action. The working feature is the outward deflection of the flexible cover. In rigid metal or plastic rollers, there is no deflection and no wrinkle-removing action. Grooves in rigid rollers are there as tread to allow air to escape from between the web and roller to increase traction.
At a vendor inspection, my tour guide – the vendor’s process manager – pointed out the vendor’s rigid, metal, spiral-grooved rollers, calling them wrinkle-removal rollers. These were obviously air-entrainment rollers for the vendor’s glossy plastic film. This left me wondering about the vendor’s web-handling knowledge and troubleshooting skills. Knowing how air entrainment and wrinkle removal work, and about different roller designs, can help you pick the right kind of roller to solve a wrinkle problem or an air-entrainment problem.
Conclusion
As a web-handling trainer and consultant, my goal is to have people not make these mistakes. There is no substitute for having trained, willing and empowered employees to avoid and solve web-handling problems. Anything less is a mistake.

Ron Lynch, president/owner of R J Lynch & Associates, holds a Bachelor of Science in Mechanical Engineering from the University of Cincinnati and is a registered Professional Engineer. He worked 39 years for Procter and Gamble, with 36 of those years working with web products and web-handling. In the early 1990s, Ron became the in-house web-handling trainer and consultant for P&G’s web-based products and its equipment and web suppliers. He has conducted more than 200 classes and trained over 3,000 engineers and technicians on web-handling in 10 countries on four continents. He was the principal instructor for the Basic Web Handling Seminar at the Web Handling Research Center at Oklahoma State University from 2000-2019. Today, R J Lynch & Associates provides web-handling training and consulting services for clients. Ron can be reached at 513-503-5160, email: [email protected].