By Ronen Dolev | 2026
A Failure Nobody Could Explain
At a flexible-packaging plant, an unexpected print-quality defect began appearing during digital printing using Liquid Electrophotography (LEP). The press appeared to operate normally, color quality met requirements, and calibration procedures revealed no significant deviation. Yet, during production, areas occasionally appeared where the ink layer had not transferred completely to the substrate.
Sometimes the defect was almost invisible. At other times, pale areas interrupted solid color fields, leading to rejected material and production stoppages.
The behavior was inconsistent. Some rolls printed without incident, while others showed intermittent defects, even though all had been supplied under the same material specification and used the same substrate definition.
The production team followed the established troubleshooting routine. They performed checks and calibrations, inspected printing-system components, reviewed image-transfer conditions, and verified the substrate settings. Attention then turned to the intermediate transfer surface and the image-forming system.
Some interventions brought temporary improvement. After one service intervention, several runs were acceptable, and the team believed the problem had finally been resolved.
Then the defect returned.
As the attempts continued, the assumption that a printing-system component was responsible became increasingly persuasive. The reasoning seemed straightforward: if ink was not transferring completely to the substrate, the likely place to look was the system responsible for transferring it.
Production hours were lost, material was rejected, and customer pressure increased. Despite everything that had been done, nobody could explain why the defect disappeared after certain interventions only to return later.
At that point, an application expert from the press manufacturer arrived at the plant.
The Question That Changed the Investigation
The expert asked to review the rejected prints, the test results, and the full sequence of actions taken before he arrived. He spoke with the operators, maintenance personnel, and the quality engineer. He wanted to understand not only what had been done, but why each action was chosen and what changed afterward.
After reviewing the evidence, he asked a simple question:
"Why are you so certain the problem begins in the printing system?"
One of the operators pointed at the samples. "Because the ink isn't transferring properly. You can see it."
"That's true," the expert replied. "We know what the result looks like. But do we know what caused it?"
That question changed the discussion.
Until then, the team had focused on identifying which component was not functioning properly. Now it had to consider another possibility: perhaps the printing system was responding to an unstable condition somewhere else in the process.
This was not a matter of semantics. It was the difference between identifying a symptom and establishing a root cause.
When a print defect appears, the natural reaction is to look for its origin in the mechanism that produces the visible result. When color becomes unstable, we check calibration. When ink transfer is uneven, we inspect the image-transfer system. When marks repeat, we examine rotating components.
These are sound, professional starting points. They are not always sufficient.
The problem begins when the most plausible explanation becomes the only explanation a team is willing to consider, long before the evidence has proved it.
Understand the Technology Without Becoming Trapped by It
In LEP printing, image formation and transfer rely on an electrophotographic process in which electrically charged particles of liquid ink develop the image. The developed image is transferred to an intermediate surface, where the ink undergoes a thermal process before being transferred to the substrate.
The final result depends on an interaction of electrical, thermal, and mechanical processes. Image formation, ink development, the condition of the intermediate transfer surface, the contact conditions at transfer, and the substrate's behavior must all operate within a defined process window.
A transfer defect therefore does not, by itself, prove that the cause lies in the image-forming or transfer system. Similar visible symptoms can arise from very different mechanisms.
The expert asked the team to stop making further parameter changes temporarily and return the press to a documented, approved baseline.
This did not mean the earlier actions were necessarily wrong. It meant that another change, made without a clearly defined hypothesis, might make the real cause even harder to identify.
Before trying to correct a problem, you must know exactly what you are trying to explain.
The First Clue Did Not Fit the Accepted Explanation
The expert asked the team to arrange the rejected samples in production order and examine each defect's position.
Several samples showed pale areas on the same side of the printed web. At first, this seemed to reinforce the suspicion that a particular component in the image-transfer system was responsible.
But measuring the physical distances between the defects along the web told a more complicated story.
The defects looked similar, yet they did not occur at a fixed repeat interval. Their spacing could not be consistently matched to the circumference or operating cycle of any particular rotating component.
That finding did not eliminate every possible transfer-system fault. It did, however, weaken the hypothesis of one permanent, localized defect on a rotating component.
One engineer proposed replacing the transfer blanket again. "The last time we changed it, print quality improved significantly," he explained.
"What else changed during that replacement?" the expert asked.
The engineer explained that the press had been stopped, the component replaced, the required procedures completed, and production restarted.
When they checked the production records, another detail emerged. The substrate roll had also been changed during the same stop because the previous roll had run out.
"So you didn't change only the component," the expert said. "You also changed the material you were printing on."
Suddenly, the improvement that had seemed to confirm a transfer-blanket problem was no longer conclusive.
For the first time, the team looked at that service intervention not simply as maintenance, but as an experiment in which more than one variable had changed.
A temporary disappearance of a defect after an intervention does not prove that the intervention removed the root cause. Sometimes it merely changes the conditions under which the problem becomes visible.
The First Experiment: Was the Substrate Changing the Result?
To narrow down the possibilities, the team designed a comparative print test.
They selected a test image containing solid color areas, different ink-coverage levels, and details that made interruptions in transfer easier to identify.
First, they printed it on a reference substrate that had already demonstrated stable performance on the press. The result was acceptable.
Next, they printed on material from the group of rolls associated with the failures, using the approved operating settings. The print initially looked good. Later in the run, the defect reappeared.
One participant concluded that the substrate must be the cause.
The expert was not ready to agree. "We've shown that the result changes between the rolls," he said. "We haven't yet shown whether the difference lies in the surface, the structure of the material, or the way it travels through the press."
That distinction matters, especially when printing on polymeric films.
Two rolls may have the same commercial name, be manufactured from the same raw-material family, and meet the same specification, yet behave differently in thickness uniformity, local elongation, flatness, or winding characteristics.
An acceptable result on one roll does not prove that another roll is defective. It tells us that we need to understand what distinguishes them.
The team also changed the position of the solid color areas within the test image. The aim was to determine whether the defect followed a particular digital image element or remained associated with a physical location along the material path.
The defect continued to appear in a similar cross-web region. The hypothesis of an image-file-related fault became less convincing.
But a satisfactory explanation was still missing.
The Observation That Never Made It Into the Report
During another discussion around the samples, one of the operators made an observation almost in passing.
"There's something we haven't checked. Almost every time we saw the problem, the roll didn't seem to be equally taut across its width."
The expert asked him to explain.
The operator recalled that, on some rolls, a narrow area of the web appeared slightly slacker than the rest. The difference was not obvious, and it did not always produce a visible wrinkle.
He had not considered it important because the press's tension-control system showed no abnormal reading.
The expert asked to see one of the suspect rolls.
They inspected the material before it entered the printing section. At first glance, everything looked normal: no major wrinkles, tears, or obvious visual damage.
However, a closer inspection and controlled off-press flatness tests indicated that a particular cross-web region did not behave like the rest of the film.
The expert then asked the team to review the cross-web thickness profile and the winding condition.
The results indicated small differences across the material and some nonuniformity in the roll's stiffness profile. None of these findings, in isolation, necessarily justified rejecting the film.
Together, however, they raised a different question.
What if the problem was not how the image was formed, but whether the substrate maintained consistent local flatness as it passed through the ink-transfer zone?
What a Material Specification Does Not Always Tell Us
Uniformity is an important requirement in polymer-film production, but no manufacturing process produces material with perfectly identical properties at every point.
Small variations in thickness, elongation, and local strain can influence how film is wound, stored, and unwound.
When thickness variations persist in the same cross-web region over many layers, they can produce localized buildup within a roll and differences in winding hardness.
Nonuniform elongation across the web can also produce relatively slack areas, a condition often described in the converting industry as a baggy web.
Such a web may travel through the handling system even while the average web tension remains within the approved range. Measuring total web tension does not necessarily describe how that tension is distributed locally across the material width.
If one band of film is slightly longer or less taut than its neighbors, local flatness may be compromised. Under some conditions, the material may also become more prone to wrinkling as it passes over rollers or through contact zones.
In LEP printing, the ink image is transferred from an intermediate surface to the substrate through a physical contact process. A local disturbance in contact conditions is therefore one plausible mechanism for an interruption in transfer.
Of course, a web that is not perfectly flat does not, on its own, prove the origin of a print defect.
A defensible conclusion requires evidence connecting the location of the material irregularity, its physical behavior, and the printed result.
That became the next task.
The Experiment That Opened the Way to a Solution
The team designed a set of tests to distinguish among three possibilities: a failure in image formation, a difference in the substrate's surface condition, and a mechanical influence of web flatness on ink transfer.
They began with two rolls of the same substrate grade: one that had performed consistently in previous runs, and one associated with rejected prints.
Surface checks, including assessments relevant to the treatment condition and the printing interface, revealed no clear difference that could account for the phenomenon. This did not completely rule out localized chemical variation, but it shifted the focus toward another explanation.
The flatness tests, by contrast, produced reproducible differences.
The acceptable roll behaved relatively uniformly across its width. The suspect roll contained a localized region that tended more toward slackness and distortion under defined mechanical conditions.
The team compared this region's cross-web position with the locations of the print defects and found a recurring match.
They then tested additional rolls from the same delivery, including rolls originating from different cross-web positions in the parent roll. Throughout the comparison, they preserved approved print settings and used the same print-quality criteria.
Rolls with acceptable flatness profiles produced stable prints. On rolls showing the mechanical nonuniformity, the defect returned in a corresponding cross-web region.
Finally, the team received a replacement roll whose thickness profile and flatness had been checked in advance. The test image and a representative production run were repeated.
This time, ink transfer remained uniform throughout the run.
The findings supported the conclusion that the substrate's flatness and mechanical behavior were the principal contributors to the defect under the conditions tested. The likely mechanism was a localized disturbance in the contact conditions between the web and the intermediate transfer surface.
However, without directly measuring contact in the transfer zone at the moment of failure, the team could not establish every detail of that physical mechanism with certainty. The corrective action therefore included monitoring subsequent runs and reviewing the findings with the substrate supplier.
The solution was not to replace another printing component. It was to prevent rolls with unsuitable mechanical behavior from entering the application while investigating the source of variation in film production and winding.
Why the Earlier Fixes Seemed to Work
After completing the investigation, the team returned to the earlier troubleshooting record.
They discovered that several interventions followed by reported print-quality improvements had also involved loading a different substrate roll.
Because the team had already become convinced that the defect originated in the image-transfer system, those improvements had been attributed to the component that was replaced or serviced.
The new tests demonstrated that changing the roll was a significant variable that had never been isolated.
The problem was not that maintenance personnel had performed service work. The problem was that the positive outcome had been interpreted as proof of the root cause without examining the other conditions that changed at the same time.
This is a familiar trap in industrial troubleshooting: confusing an action that precedes an improvement with an action that actually causes it.
The trap becomes harder to escape when the improvement appears to support what the team already believes.
The more time, effort, and resources you invest in one hypothesis, the harder it can be to accept that the investigation may have been moving in the wrong direction.
Engineering judgment is not only the ability to propose another solution. It is also the willingness to recognize when the existing explanation no longer fits the evidence.
Thinking Outside the Box Is Not Searching for a Magic Fix
The expression thinking outside the box is often used to describe creative, unusual, or surprising solutions. In industrial printing, it means something more demanding.
It does not mean ignoring established knowledge, manufacturer instructions, or the principles of physics and chemistry. Nor does it mean trying random parameter changes and hoping one of them works.
Quite the opposite.
To think differently, you first need a strong understanding of how the system is supposed to operate. Only then can you recognize observations that do not agree with the expected behavior.
In this investigation, understanding LEP image transfer was essential at the outset. But that knowledge alone was not enough to resolve the failure.
The team had to look beyond the mechanism where the defect became visible and examine the material's history and mechanical behavior during printing.
Progress did not come because someone knew the answer in advance. It came because someone was willing to ask a different question.
Thinking outside the box often begins with a detail that refuses to fit: an irregularity that does not repeat as expected, a temporary improvement without a satisfactory explanation, a reading that is technically acceptable but does not describe the entire phenomenon, or an operator's observation that never found its way into the formal report.
Such clues do not prove anything on their own. Their value is that they make it possible to form a new hypothesis and design an experiment to test it.
Even an Expert Can Be Wrong — and That Is Not a Professional Weakness
During a difficult investigation, people often expect the most experienced person in the room to recognize the solution quickly. But experience does not eliminate uncertainty.
Experience helps identify plausible mechanisms, choose meaningful tests, and interpret the results. It also helps us recognize when a conclusion goes beyond what the evidence actually supports.
Yet experience can become a limitation when it creates confidence too early.
A specialist who has seen dozens of similar failures may quickly recognize a familiar pattern. That is often an excellent starting point. But treating the current case as identical to earlier ones can conceal the small difference that changes everything.
One of the most important disciplines in troubleshooting is therefore separating experience from proof.
Experience can suggest where to look. The experiment must establish whether the direction is right.
A good hypothesis is not one that cannot be challenged. It makes a testable prediction and can be revised when the results disagree.
If the evidence contradicts a hypothesis, the investigation has not failed. An explanation has been weakened or eliminated, and the team's understanding has improved.
Sometimes the greatest progress comes at the precise moment we realize that the explanation we initially found so convincing cannot account for all the facts.
Turning This Approach Into a Practical Method
The lesson is not that every ink-transfer problem comes from web flatness or winding quality. That would repeat exactly the mistake this article is trying to prevent.
The lesson is how to investigate when the root cause is not known.
First, separate observation from interpretation. An area where the ink layer is discontinuous is an observation. Calling the transfer blanket defective is a hypothesis until evidence supports it.
Second, examine what changed before the failure appeared. Look not only at press parameters, but also at the substrate, material lot, winding condition, storage history, and processes that preceded printing. The relevant change may lie outside the area the team habitually searches.
Third, actively look for evidence that challenges the favored explanation. If a rotating component is suspected, test whether the position and repeat distance fit that hypothesis. If the substrate is suspected, determine whether the rejected rolls genuinely differ in a property that has a plausible failure mechanism.
Fourth, design experiments that distinguish competing hypotheses rather than merely produce better-looking samples. Changing several parameters at once can improve the result while making it nearly impossible to identify which change mattered.
Fifth, listen to the people who work with the process every day. Operators and maintenance personnel may notice behavior that doesn't show up on a dashboard. Their observations don't replace measurement, but they can point the team to the right question and the right test.
Finally, validate the corrective action under real production conditions. One acceptable sample is not enough. Recreate the circumstances associated with the defect, demonstrate process stability over an appropriate run, and verify that the corrective action has not introduced another failure.
The Cost of an Incorrect Diagnosis
In an industrial plant, every technical decision affects time, materials, costs, and customer commitments.
Replacing components that are not responsible increases maintenance cost. Repeated adjustments consume production time and substrate. Unplanned stoppages reduce press utilization, and poorly controlled tests can introduce further variation.
But the highest cost is sometimes the defect that keeps returning because no one ever understood it.
At the same time, an investigation that continues indefinitely without interim decisions can be expensive. The team must define which production can safely continue, which material should be isolated, which tests are necessary, and what evidence will be sufficient to authorize a return to normal operation.
The objective is not to achieve a perfect theoretical explanation at any cost. It is to understand the failure well enough to define a corrective action, demonstrate its effectiveness, and maintain a stable, reproducible process.
Done properly, an investigation does more than resolve the immediate failure. It creates knowledge the plant can use when the next unexpected problem appears.
Rejected hypotheses belong in that knowledge record, too. They explain what was examined, why an explanation was discarded, and what the team learned along the way.
Smarter Presses Still Need Better Questions
Modern digital presses incorporate sophisticated control systems, sensors, automated calibration, and continuous monitoring of many process variables.
These technologies improve consistency and make it easier to detect deviations, but they do not ensure that every possible failure mechanism is measured directly.
A system can display readings within specification while failing to reveal an unexpected interaction among several variables.
Sometimes the missing element is not more data, but better interpretation.
An average value does not necessarily describe local behavior. A result within tolerance does not prove that every process stage is stable. A correlation between two variables does not necessarily establish that one caused the other.
Even as control systems become more sophisticated, engineering judgment is needed to decide what the data prove, what they do not prove, and which additional test is required.
Instrumentation supplies the data. Engineering reasoning gives those data meaning.
Conclusion: Sometimes the Answer Lies in the Question Nobody Asked
By the end of the investigation, the press had returned to stable production under verified conditions. The defect that had triggered repeated checks, adjustments, and interventions in the image-transfer system did not require another replacement of a printing component.
The principal cause was traced to mechanical nonuniformity in the substrate, reflected in web flatness and behavior as the material passed through the printing area.
But the most important outcome was not simply the disappearance of the defect.
The team learned to look at failures differently.
Instead of immediately asking which parameter to change, it began asking what evidence was missing. Instead of deciding on a cause in advance, it learned to distinguish among symptoms, hypotheses, and mechanisms. Instead of treating temporary improvement as proof, it began demanding a result that could be reproduced under production conditions.
That is what thinking outside the box really means in industry.
It is not the ability to invent a surprising solution every time. It is the willingness to question our assumptions, listen to evidence that challenges them, and investigate possibilities that were not part of the original plan.
Professional experience is an enormous asset, but it is not a substitute for curiosity. Sometimes an intimate knowledge of one technology leads us to search in the familiar place, while the real answer lies at the interface with another system.
Experience tells us where to start looking. Curiosity makes us look elsewhere. Evidence tells us whether we were right.
And in industrial printing, where every decision affects quality, material, time, and money, that ability matters every bit as much as knowing the press itself.
Author's note: The situation described in this article is an illustrative engineering scenario based on established principles of LEP printing, polymer-web behavior, process control, and industrial failure analysis. It is not presented as a record of a specific customer investigation. Its purpose is to demonstrate the reasoning, diagnosis, and validation required when the root cause is not known in advance.