Canmaking AI Knowledge Hub

Question 1

Why do coating thickness readings vary from spot to spot — on a formed can or across flat sheet?

Accepted Answer

Because three different things are varying at once: the film the coater laid down, the metal underneath it, and the measurement itself. Separating them is the whole job, and the order matters — confirm the measurement, then look at geometry, then at the coater. The measurement. Before anything else, check that the readings are comparable. Correct substrate setting, clean and dry surface, probe seated square and at consistent pressure, and the same defined positions every time. On a curved surface a probe that tilts reads differently from one that sits flat. Any method that averages over the area it covers reports a mean film over that spot, not the thinnest point inside it. A measurement pattern that is not written down is not a measurement pattern. Across flat sheet the variation usually comes from the coater and the metal reaching it. Look at roller condition and pressure across the width, coating viscosity and temperature drifting through the shift, and pick-up at the applicator. Sheet flatness, camber and oven loading do the rest. The signature is positional — the same place on every sheet reads the same way. Across-web variation, front-to-back variation and sheet-to-sheet variation each point at a different part of the line. On a formed can geometry joins in. Drawing and wall-ironing thin the metal along the wall, and the film follows what the metal does. Radii, the dome, the neck and bead roots stretch the coating; compressed areas can thicken or wrinkle it. On a three-piece body the film over the plain margin, the weld area and the panel was never the same film to begin with. So expect a spread of readings around a can, and let the specification say where on the can it applies. How to tell them apart. Map before you adjust. Take readings at fixed positions, on several cans, over time. If the pattern repeats at the same positions, it is application or geometry. Movement from can to can points at handling, surface condition and the measurement method. If it drifts over hours, look at viscosity, temperature and roller wear. Most spot-to-spot disputes end at either the measurement pattern or the geometry, not at the coater.

Question 2

Why do missing lacquer defects slip through on fast lines even with visual inspection?

Accepted Answer

Because the defect is intermittent, sits inside the can, and has almost no visual contrast. Those three properties defeat human inspection, however careful the inspector. It arrives in bursts, not as a trend. A blocked stripe nozzle, a starved supply or a start-up after a stop produces a run of affected cans that may last seconds and then stop. A sampling schedule that pulls cans every half hour has no chance of coinciding with it. Inspection frequency suits characteristics that drift. This one does not. It is in the wrong place to see. Internal coating and the internal side seam stripe are inside a cylinder. Seeing them at all needs light at the right angle and an unobstructed view down the can, which is not what a person walking a running line has. There is nothing to see even when you look. Many internal coatings are near-transparent or pale gold on bright metal. A missing patch does not appear as a mark — it appears as slightly different reflectance, and only under the right light. Clear stripe over a weld is harder still. Rare defects get missed more, not less. Sustained visual inspection of a repetitive task has a miss rate that rises with time on task and with fatigue. It rises further when defects are rare. An inspector who has not seen a fault all shift is measurably less likely to spot the one that appears. This is a well-documented property of human visual search, not a comment on any individual. And the consequence is asymmetric. A missed can is not a cosmetic escape. It is bare metal in contact with food or beverage, so this defect normally gets 100 percent in-line detection rather than visual sampling. Visual inspection stays valuable. It confirms what a detector rejected and explains a defect once someone finds one. It is simply the wrong instrument for catching it in the first place.

Question 3

What is the actual cost of a missed coating defect that reaches the seamer?

Accepted Answer

There is no universal figure, and anyone quoting one is guessing at your plant. But the shape of the cost is consistent, and it escalates sharply with every station the defect passes. The cost comes in stages, not as one number. Caught at the coater, the loss is metal and a little line time. After forming, it is a can, plus everything added to it since the coater. Caught at or after the seamer, the can is finished goods. It already carries the value of the body, the end, the forming, the energy and the labour. Past the filler, the product, the closure, the filling and any thermal process are in it too, and the pack now belongs to a customer. The escape is never one can. This is the part that surprises people. When a defect turns up downstream you rarely know when it started. So the decision is not about that can. It covers everything made since the last known-good verification. That means quarantine, sorting or scrapping a batch whose size your verification interval sets, not the size of the fault. A longer interval between checks is a larger batch at risk. Then the indirect costs. Line stoppage and re-qualification. The investigation itself. Re-inspection labour. If it reached the filler, their complaint process, sorting, returns and credits — plus the standing you lose as a supplier. In the worst case, product in the market with corrosion, perforation or taint, which is recall territory for the filler and reputational exposure for whoever made the can. How to work out your own number. Take the value of a finished can at the seamer. Multiply it by the number of cans the line makes between two verifications of the coating check. That is the exposure of a single undetected event, before any downstream cost. Then ask how often such an event actually occurs on your line. That calculation is worth doing properly, with your own scrap rates, line speed, product value and customer agreements. The Innosen ROI calculator walks through the same arithmetic if the structure helps.

Question 4

What is Gauge R&R, and why does it matter for coating equipment?

Accepted Answer

Gauge repeatability and reproducibility is a study that separates the variation coming from your measurement system from the variation coming from your process. Repeatability is the variation when the same person measures the same can with the same gauge more than once. Reproducibility is the variation between different people measuring the same cans. The output tells you how much of the variation you are charting is real and how much you invented by measuring. If the measurement system consumes a large share of the tolerance, control charts react to noise and capability indices understate the process. Good product then fails while bad product passes. The widely used AIAG measurement systems analysis guidance sets three bands. Under ten percent of study variation is acceptable. Ten to thirty percent is conditionally acceptable, depending on the application and the cost of improvement. Above thirty percent is unacceptable. Confirm which criterion your own quality system and your customers require before you quote a figure. Why it bites particularly hard on coating. Coating specifications are often narrow relative to what any measurement system can resolve. The measurement is also indirect: it infers a film from an electrical or magnetic property, on a curved surface, through a probe that an operator seats by hand. That combination makes it easy for the measurement to eat a large part of the tolerance without anyone noticing. When it does, two expensive things follow: you chase the coater for drift that was never there, and you miss drift that was. A Gauge R&R study is what tells you which of the two you have been doing. The same logic extends beyond coating thickness. Enamel raters have a method and an electrolyte that vary between operators. In-line detection systems make pass or fail judgements rather than measurements, so they call for an attribute agreement study rather than a classic Gauge R&R. The question stays identical: can this system tell good from bad reliably enough to act on?

Question 5

What are the most commonly missed defects in enamel rating, and why do they get through?

Accepted Answer

Enamel rating reports one number for total exposed metal. So anything that adds little current, sits outside the wetted area, or hides behind the fixture will pass a test somebody ran correctly. A low reading means no evidence of exposure — not no exposure. Defects outside the wetted area. The electrolyte only tests what it touches. Exposure above the fill line, on the flange, on the curl, on the seaming panel or under the seal of the adapter contributes nothing to the reading. Component fixtures exist precisely to define that boundary, and using the wrong adapter moves it. Small exposures in critical places. A fine crack along an easy-open score, or exposure at the rivet where the metal travelled furthest, may add very little current. Yet that is exactly where an acidic product attacks first. Area and consequence are not the same thing, and enamel rating measures area. Film that is continuous but weak. A thin, under-cured or porous-but-unbroken coating conducts little today and fails later — after retort, after months of storage, or once the product has had time to work on it. The test sees a barrier; it cannot see how long that barrier will last. Defects masked by test conditions. Short dwell, incomplete wetting of a bead root or rib, and air bubbles trapped against the surface all suppress the reading. So does electrolyte that is cold or weak. These produce false confidence rather than a wrong number, which is harder to notice. Defects averaged away. Where a report gives a batch mean, one bad can among several good ones disappears. Individual values and their spread carry information the average destroys. Damage created after the sample point. Handling, transfer and forming strip coating off downstream of wherever you take samples. A body that rates clean before flanging and beading tells you nothing about the finished can. What closes the gaps. Watch where the bubbles form during the test rather than only reading the meter — position separates an application fault from a handling one. Fix the dwell, fill level and wetting method so results are comparable. Test at more than one point in the process, including the finished can. Use the right fixture for the component. And pair the electrical result with a visual examination of anything out of the ordinary, because the number alone will not tell you what you are looking at.

Question 6

Lacquer, enamel, coating, varnish — what is the difference?

Accepted Answer

In canmaking these words all mean the same thing: an organic coating that goes onto metal and then cures. Regional habit decides which one you hear. European plants tend to say lacquer or varnish. North American plants tend to say enamel or coating. One conversation can call the internal protective layer on a food can the internal lacquer, the interior enamel and the food-contact coating. The word does carry a hint of function. Varnish usually means a clear external coating over print or bare metal. Enamel most often refers to the internal protective film — hence enamel rating for the test that checks it. Size or primer refers to a base coat under print. When a specification matters, read what the coating has to do rather than what someone calls it.

Question 7

Why does the substrate — tinplate, TFS or aluminum — change how coating is measured?

Accepted Answer

Because most coating measurement methods measure the coating indirectly, through its effect on an electrical or magnetic property of the substrate underneath it. Change the substrate and you change the signal. Tinplate is steel with a tin layer; TFS, also called ECCS, is steel with a chromium and chromium-oxide layer; aluminum is non-ferrous. Ferrous and non-ferrous substrates call for different measurement principles. Surface treatments such as passivation add their own thin layers, which the method may or may not include in what it reports. The practical rules are the same in every plant. Calibrate on the substrate you are actually running. Treat a substrate change as a change that requires re-verification. Never compare readings from different substrates as if they were one measurement. Substrate also changes the coating's job, because TFS, tinplate and aluminum do not share the same barrier requirements. So a coating specification belongs to one substrate and one product.

Question 8

What is a DWI can, and what is a DRD can?

Accepted Answer

DWI means drawn and wall-ironed; DRD means draw-redraw. Both are two-piece. The difference is how far the process thins the metal. In DWI the press draws the cup, then pushes it through ironing rings that thin and lengthen the wall considerably. That gives the tall, thin-walled body typical of beverage cans. In DRD the press draws and redraws the blank into a shallower shape, keeping most of the wall thickness. That suits food cans, which have to survive processing and handling. From a coating standpoint, DWI creates a wall whose thickness and surface condition change along its height, plus a strongly deformed dome. DRD thins less, but leaves sharp radii in the draw corners. Both take their internal coating after forming, and both need a coating continuity check.

Question 9

What is the difference between a two-piece and a three-piece can?

Accepted Answer

A two-piece can is a body and base formed from one disc of metal, closed with a single end. A three-piece can is a rolled body with a welded side seam, closed with two ends. The difference determines where coating goes on and where it is at risk. On a two-piece can the coating goes on after forming, because forming would destroy a pre-applied film. The press draws the body, or draws and wall-irons it, and the line then spray-coats and cures the interior. The coating question is coverage over a continuously varying wall thickness and over the dome and neck. On a three-piece can the flat metal takes its coating first, usually on a sheet-fed coater. Cutting, welding, striping, flanging, beading and seaming all follow. The coating question is different: plain margin placement, coating survival through forming, and the integrity of the side seam stripe over the weld.

Question 10

How are inverted or misplaced blanks detected?

Accepted Answer

By checking a property that differs between the two faces of the blank — most often the presence, type or reflectivity of the coating and print on one side. A blank carrying coating on one face only, or print on one face only, presents a different surface to a sensor depending on which way up it sits. The reason it matters is straightforward. An inverted body blank puts the external coating inside the can and the internal coating outside. That is a product-contact failure, not a cosmetic one. The same logic applies to inverted ends and to labels or decoration applied the wrong way round. These checks sit as close to the feeder as possible, so the wrong blank never reaches forming.

Question 11

Sheet skew, registration and translation — what is the difference?

Accepted Answer

Three different errors in how a sheet sits: skew is rotation, translation is sideways displacement, and registration is the position of one applied layer relative to another. Measure them separately, because they have separate causes and separate fixes. Skew is angular. The sheet is square but presented at an angle, so the error grows from one side of the sheet to the other. Translation is linear and lateral. The sheet sits square but too far to one side, so the whole image, the coating and the plain margin all shift across the sheet by the same amount. On a coil line the equivalent is the strip wandering across the rollers. Registration is relational. It asks whether the layer going on now lines up with the layer already there. Color to color in printing, varnish to print, coating to blank edge, or the printed design to the line where the slitter will cut. A sheet can sit perfectly square and still print out of register, because an earlier station did not. Keeping them apart has practical value. A margin wrong by the same amount everywhere is translation. Where the error grows from one edge to the other, that is skew. A margin that is correct but printed over is registration. The defect looks similar on the finished blank and the three causes are in three different places.

Question 12

What is sheet skew, and why does it have to be measured?

Accepted Answer

Skew is angular misalignment. The sheet arrives rotated against the direction it should travel, or against the machine's reference edge. A skewed sheet is not a jam. It usually runs through perfectly well, which is exactly why you have to measure it rather than watch for it. The consequences appear one station later. Coating and print land at an angle to the sheet edge, so the plain margin narrows at one corner and widens at the other across the same sheet. Blanks slit from that sheet are out of square, which puts the weld overlap under tolerance at one end. Print register drifts between colors. On end and component presses a skewed blank loads unevenly into the tooling. Skew is progressive. It comes from feeder side guides, worn grippers, an out-of-square stack, uneven suction or belt tension. So it drifts through a shift rather than appearing all at once. Measuring it continuously at the infeed catches the trend while it is still a setting and not a batch of scrap.

Question 13

What is measured on a coil and shear line?

Accepted Answer

Dimension and position, before the metal reaches a coater. Coil and shear lines uncoil the strip, cut it to sheet and stack it, and everything downstream inherits whatever comes off that stack. Five characteristics matter. First, the lateral position of the strip as it runs: coil translation, or steering. A strip that wanders puts the shear cut in the wrong place across the width. Second, sheet length after shearing, taken on-line rather than by pulling sheets out of the stack. Third, sheet width and squareness, because no feeder can register an out-of-square sheet later. Fourth, edge and burr condition, which affects both handling and the eventual weld. Finally, flatness or camber, which decides how the sheet lies on the coater and in the oven. None of these are coating measurements. Still, every coating and registration problem further down the line can start here. So the first question in a print-register or plain-margin investigation is whether the sheets themselves are square and the right size.

Question 14

What is double sheet detection and why does a press need it?

Accepted Answer

Double sheet detection identifies when two sheets or blanks are fed together instead of one, and stops or diverts them before they enter the tooling. A press or coater fed a double thickness can damage dies, rolls and feed mechanisms, and the repair costs far more than the sheet. It sits at sheet feeders into coaters and printers, at blank feeders into presses making ends and components, and anywhere a destacker runs. The related checks are for misfed, skewed, inverted or missing sheets, because the same feeder that occasionally takes two also occasionally takes none, or takes one the wrong way up. Detection settings have to tolerate normal variation in plate thickness and in the oil film on the metal without missing a genuine double.

Question 15

What goes wrong when too much lacquer is applied?

Accepted Answer

Treat over-application as a defect, not as extra safety margin. A heavy film does not cure the way a correctly applied one does. Solvent has further to travel out of the film. So under-cure, softness and solvent retention all become more likely at the same oven setting. Thick film is also less formable. Coating that looks fine on a flat sheet can crack or flake once the metal goes through drawing, ironing, flanging, beading or curling. That leaves bare metal in exactly the places that are hardest to protect. Add the cost of the coating itself, blocking and sticking problems in stacks, and print and overvarnish appearance issues. The case for controlling the upper limit is then as strong as the case for the lower one.

Question 16

What is a coating skip, and how does it differ from a pinhole?

Accepted Answer

A skip is an area the coating never reached. A pinhole is a small hole through coating that did go on. The distinction matters because they point at different parts of the process. Skips come from the application step: roller wear or damage, insufficient pick-up, a gap in the film as it transfers, or metal that did not present flat. Pinholes come from what happened in or after the film. Entrapped air or solvent escaping during cure. Contamination or dirt particles in the wet film. A foreign body burning out in the oven. Or a defect in the metal surface pulling the film apart as it flows. Both show as exposed metal in an enamel rating test, so the electrical result on its own will not tell you which you have. Visual inspection of the defect, and where on the can it appears, will.

Question 17

What is plain margin inspection?

Accepted Answer

Plain margin inspection checks that the uncoated strip left for welding is present, wide enough, and correctly positioned on the coated sheet. It is a coating-placement check rather than a coating-quantity check. It catches four faults. Coating that has run into the margin. A margin that has drifted off position across the sheet. A margin that has grown too wide because the coater is under-covering. And skew between the margin and the sheet edge. Each of those causes a problem downstream. A contaminated margin welds badly. Too wide a margin leaves more bare metal for the internal stripe to cover. A skewed margin produces a weld that wanders relative to the coating edge.

Question 18

What is plain margin, and why is it left uncoated?

Accepted Answer

Plain margin is the deliberately uncoated strip along the edge of a coated sheet or blank. It stays bare so the welder has clean metal. Coating is an insulator; resistance welding needs metal-to-metal electrical contact. If lacquer runs into the margin, weld quality suffers. If the margin runs too wide, the internal stripe has more bare metal to cover than its setting allows. Plain margin is therefore a dimensional quality characteristic in its own right. Its width, its position relative to the blank edge, and its consistency along the sheet all matter. Check it on the coated sheet before the blanks go to the welder, because a margin fault found at the welder is already a sheet-level scrap decision.

Question 19

What is coating weight, and how is it different from coating thickness?

Accepted Answer

Coating weight is the mass of dry coating per unit area, typically milligrams per square inch or grams per square meter. Coating thickness is the physical depth of the film, in microns. They describe the same layer in two ways and are related through the density of the cured coating. Each specification uses whichever unit the coating supplier and the canmaker agreed. Converting between them requires a reliable density figure for the cured film, so a value converted from one unit to the other is only as good as that density assumption. Where a specification uses film weight, measure and report in film weight.

Question 20

What is coating thickness measurement in canmaking, and why does it matter?

Accepted Answer

Coating thickness measurement tells you how much lacquer or enamel sits on a defined area of the metal. Reports usually give microns of dry film, or film weight per unit area. On a can, the coating is the barrier between an acidic or corrosive product and the tinplate, TFS or aluminum underneath it. Too little coating and the barrier is incomplete: the product attacks the substrate, the can corrodes, and metal pick-up or perforation follows. Too much coating costs material, slows curing, and can crack once the metal goes through drawing, flanging or curling. The specification therefore has a lower limit set by product protection and an upper limit set by cost and formability, and measurement exists to prove you are inside both.

Question 21

What is checked at printing and varnishing?

Accepted Answer

That the press is decorating the right face, that the layers line up, and that decoration has not covered something that must stay bare. Printing is an external operation, so it rarely touches product safety directly — with two exceptions that do. The first is the plain margin. Print or overvarnish that creeps into the margin insulates the metal and the weld suffers, exactly as coating in the margin does. The second is sheet orientation: an inverted sheet puts the decoration on the face that should carry the internal coating, which is a product-contact fault rather than a cosmetic one. Alongside those, the routine checks cover color and print register between stations, varnish coverage and film weight, and print-to-slit position so each blank carries a complete design. On the finished can, check that the label or decoration sits the right way up. Inverted can label detection exists because a can that is upside down relative to its print is not detectable by anything looking only at the print itself.

Question 22

Why does cure state affect coating measurement and enamel rating?

Accepted Answer

Because both methods measure a cured film, and an under-cured or over-cured one is not the film the specification assumed. The chemistry of curing and the oven schedule are outside what Canmaking AI covers; the effect on measurement is not, and it comes up constantly. An under-cured film still holds solvent. It is softer, so it marks and abrades in handling between the oven and the laboratory. It also behaves differently as a dielectric. That can shift a coating thickness reading taken shortly after the oven against the same spot read later. It is also more permeable and less continuous, so enamel rating reads higher without any change at the coater. An over-cured film is harder and more brittle. It survives handling but cracks when forming works the metal, so the exposure shows up at enamel rating on the finished can rather than on the flat sheet. The practical rule is to fix when you measure, not just how. Measure at a defined point after cure, on cans or sheets that have reached a consistent temperature. Treat a reading taken warm out of the oven as a different measurement from one taken at the end of the line. When enamel rating rises and film weight has not moved, cure is one of the first things to check — with the people who own the oven.

Question 23

What is missing lacquer detection, and where on the line does it sit?

Accepted Answer

Missing lacquer detection checks 100 percent of cans in-line for internal coating where it belongs. The system rejects uncoated or partially coated cans before they reach the filler. It exists because the defect is intermittent, invisible from outside the can, and unacceptable in food and beverage packs. On a three-piece line the check sits after internal coating and curing, and before flanging, beading or palletising. That way a fault surfaces while the cans still trace back to a coater setting. Detection of a missing side seam stripe belongs after striping and cure. Two things decide whether the installation earns its place. Does it catch the defect reliably at line speed? And does it reject good cans often enough that someone switches it off? Commissioning sets both, and you should re-verify them with known good and known defective cans on a defined schedule.

Question 24

Why does the side seam of a three-piece can need a separate stripe?

Accepted Answer

Because the coating has to come off, or stay off, before welding — and then go back on afterwards. A welded side seam needs clean, bare, conductive metal at the edges of the blank. So the coater leaves a plain margin there, or a later step strips the coating off. After welding there is a strip of unprotected metal running the full height of the can, plus the weld itself and the heat-affected zone either side of it. The side seam stripe — internal, external or both — restores the barrier over that strip, and then cures. It is one of the most defect-prone operations on a three-piece line. The application is narrow, fast and precisely timed, onto a moving can. And missing it leaves bare metal in direct contact with the product. For that reason, food and beverage lines check internal stripe coverage continuously.

Question 25

What causes missing lacquer on a food can?

Accepted Answer

Missing lacquer means an area of the can that needs coating has none, leaving bare metal in contact with the product. On a three-piece food can, look in three places. The internal side seam stripe, the area next to the weld, and patches on the body where the coating failed to transfer or came off again. The causes group into four. Application faults, where the stripe gun or roller never delivered coating at all. A blocked nozzle, an empty or starved supply, a timing error against the can, a stopped or slipping applicator. Transfer faults, where the coating went on but did not stay — contaminated or oily metal, poor wetting, the wrong surface treatment. Mechanical removal, where guides, star wheels, transfer belts or tooling scraped, wiped or abraded the coating off again. And process faults on start-up and after a stop, where the first cans through carry no stripe or only a partial one. Because the defect is intermittent by nature, sampling alone rarely finds it. This is one of the cases that calls for 100 percent in-line detection instead.

Question 26

What happens at the slitter, and what is checked there?

Accepted Answer

The slitter cuts coated and printed sheets into body blanks. It is the last point where a sheet-level fault is still cheap to catch. After this the metal is a can body, and a fault becomes a can to scrap rather than a sheet to scrap. Check blank dimensions and squareness, since the blank width sets the can diameter and the weld overlap. Check plain margin position and width on each individual blank, not just on the sheet it came from. A skewed or mistranslated sheet produces blanks whose margins differ across the stack. Check edge and burr condition, which affects both the weld and the handling. Check that the slit position matches the printed and coated image. Finally, check blank orientation, so that no inverted blank reaches the welder. The margin check matters twice over here, because the internal side seam stripe further down the line covers a margin of one expected width. A wider margin leaves more bare metal than the stripe setting covers.

Question 27

Does enamel rating on an end differ from enamel rating on a body?

Accepted Answer

The principle is identical; the fixture, the exposed area and the interpretation are not. A can body presents a large, mostly cylindrical coated area plus, on three-piece cans, the side seam. An end presents a smaller area but concentrates the severely formed features: countersink, chuck wall, curl, score and rivet. Because the exposed areas differ, body and end results are not directly comparable and are specified separately. A dedicated adapter holds the end, isolates the product-contact surface, and keeps electrolyte off the curl and the seaming panel. Wetting those edges adds current that has nothing to do with coating quality. The same applies to tabs, caps, closures, capsules and tube or aerosol components: each needs a fixture that exposes only the surface you intend to assess.

Question 28

What is measured on can ends and easy-open ends?

Accepted Answer

After forming, ends need three checks: coating integrity, the condition of the features forming created, and the presence of every component the press added. Coating integrity is quantified by enamel rating on formed ends, held in a fixture that exposes only the product-contact surface. Beyond coating, quality assurance on an end covers four more checks. The presence and correct orientation of the tab on easy-open ends. Detection of double shells or double ends in the stack. The condition of the curl and the countersink. And the presence and placement of lining compound on the seaming panel. Faults here surface in the press area and in the bagger, because nobody can rework an end with a missing tab or a fractured score.

Question 29

What coating defects are most often found on can ends?

Accepted Answer

Ends concentrate coating problems where the metal deforms most: the countersink, the chuck wall, the curl and, on easy-open ends, the score and the rivet. The coating on an end starts life as a flat coated sheet and then has to survive stamping, forming, scoring and riveting without fracturing. Typical findings: coating fracture or micro-cracking in the countersink radius and along the score line. Coating damage on the curl, from tooling or from handling in the bagger. Incomplete coverage on the rivet, where the metal travels furthest. And repair-coat or lining-compound faults on the seaming panel. Enamel rating on ends is the standard way to quantify how much bare metal all of that produced. Test formed ends rather than flat stock, because forming is what creates the exposure.

Question 30

Does the coating survive flanging, beading and necking?

Accepted Answer

A correctly applied and correctly cured film follows the metal; a film that is too thick, too brittle or under-cured does not. On a three-piece can the coating goes on flat, then has to survive every forming operation after it. So the enamel rating of a finished can almost always exceeds that of the body it came from. The places that fail are the ones where the metal works hardest. Start with the flange, where the edge turns out and stretches the coating over a tight radius. The bead roots, where the wall compresses and the film can wrinkle or fracture. And the necked area, where the diameter comes down in stages. Damage also comes from contact rather than from deformation — tooling, guides, star wheels and transfer surfaces scrape coating off cans that formed perfectly well. The way to tell these apart is to enamel rate at more than one point: on the body before forming, and on the finished can. If the body reads clean and the finished can does not, something after the coater created the exposure. The question becomes which forming or handling station did it. Tooling condition, setup and bodymaker mechanics themselves sit outside what Canmaking AI covers, but locating the stage where the coating stopped being continuous does not.

Question 31

Does a high enamel rating always mean the can will fail?

Accepted Answer

No. Enamel rating measures exposed metal, not shelf life. Whether exposure matters depends on the product, the pack, the process after filling and the intended storage life. A high reading on a can destined for a benign, dry product may be commercially irrelevant. A much lower reading on a can for an acidic or sulfur-bearing product may not be acceptable at all. What a high reading always means is that the barrier is not what the specification assumed, and that something in the process has changed. Treat it as a process signal that requires investigation and a decision against the agreed specification — not as a pass or fail verdict on the pack by itself. Questions about the shelf life of a specific product in a specific can are application questions. They need the filler, the coating supplier and the canmaker in the same conversation.

Question 32

Why does electrolyte condition affect enamel rating?

Accepted Answer

The electrolyte is part of the measuring circuit, so its conductivity is part of the reading. A salt solution conducts according to its concentration, its temperature and what has dissolved into it. Change any of those and the same can gives a different current. The practical consequences follow. Solution made up by eye rather than by weight varies between batches. Solution left open evaporates and concentrates. Warm solution conducts better than cold. And solution reused all week carries metal ions and coating residue from every can before it. Treat the electrolyte as a consumable reference material. Give it a written preparation method, a storage condition, a temperature check and a replacement rule. Record every change, because that record is the first thing you will want when a trend appears.

Question 33

Why do enamel rating results drift during a shift?

Accepted Answer

Drift within a shift usually belongs to the electrolyte or the operator, not to the coating. Electrolyte concentration changes as it evaporates, its temperature rises with ambient conditions and repeated use, and dissolved metal from tested cans accumulates in it. Each of these shifts the current for an unchanged can. Method drift is the other half. Several things move the result: dwell time creeping longer, inconsistent filling, splashing onto the plain margin or the curl, bubbles against the surface, and contact that never fully seats. A stable enamel rating program fixes electrolyte management first: a defined preparation, a defined replacement interval and a defined temperature range. It fixes the method too. Only then does within-shift movement count as a process signal.

Question 34

Why is my enamel rating increasing?

Accepted Answer

Work from the test outwards: confirm the measurement, then the sample, then the process. A rising trend is real often enough to act on, and an artefact often enough that checking the test first saves a wasted shift. The test. Electrolyte that has aged, warmed, picked up product or metal ions, or drifted in concentration reads differently from fresh solution. Check the electrolyte, the contact time, the fill level, the temperature, the cleanliness of the contact and the condition of the adapter or stand holding the can. Check that operators still follow the same method, because dwell time and wetting vary easily without anyone noticing. The sample. Cans handled, stacked or scuffed between the line and the laboratory pick up coating damage that has nothing to do with the coater. Sample at the same point, transport them protected, and test them in the same condition every time. The process. Once the test and sample are clean, work through the process. Look for a falling film weight, coater roller wear or pressure drift, or a change in coating batch or viscosity. Then look at oven temperature and line speed against cure, mechanical contact in transfer and handling, more aggressive forming, and any substrate or surface treatment change. On three-piece cans, check the side seam stripe separately from the body. A stripe problem raises the total reading while the body coating stays the same. Compare the trend against when it started and what changed at that moment. The change log usually solves an enamel rating trend, not the gauge.

Question 35

What is enamel rating, and what does the number mean?

Accepted Answer

Enamel rating measures the electrical current flowing between an electrolyte and the metal of the can. The electrolyte sits against the coated surface, with a low fixed DC voltage across them. The gauge reads the result in milliamps. A perfect, continuous coating is an insulator, so almost no current flows. Every pore, crack, skip, pinhole or scratch that exposes conductive metal to the electrolyte provides a path, and the current rises. The number is therefore a proxy for total exposed metal area, not a picture of where the exposure sits. Two cans with the same reading can carry one large defect or many tiny ones. So an out-of-specification reading calls for a visual examination first, not an immediate process change. The canmaker, the filler or the brand owner sets the specification limits according to the product going in. An aggressive product tolerates far less exposure than a benign one.

Question 36

Why do in-line detectors produce false rejects, and what causes them?

Accepted Answer

A false reject means the detector saw a difference that was real but not a defect. The usual sources are product variation nobody set the system up for, environmental change, and mechanical presentation. Typical causes fall into a short list. A new coating, color or substrate that changes the signal. Condensation, residual moisture or coolant on the can. Dust, coating mist or overspray on the sensing window. Vibration and can wobble, which changes the distance or angle between can and sensor. A change in line speed or spacing that moves the can out of the timing window. And detection thresholds set on one production variant, never re-checked when the variant changed. The correct response is to record the rejects, look at several of them, and find the common factor before loosening a threshold. A threshold widened to stop nuisance rejects will also pass a real defect.

Question 37

100 percent in-line inspection or off-line sampling — which does a coating problem need?

Accepted Answer

Use sampling for characteristics that drift and in-line inspection for characteristics that jump. Coating thickness and film weight drift: they move slowly with roller wear, viscosity and temperature, so a well-designed sampling plan with control charts detects the trend in time to correct it. Missing coating, missing stripe, inverted blanks, double sheets and similar faults do not drift. They appear suddenly, affect an unknown number of consecutive units and then stop, so a sample taken every half hour can miss the entire event. Those characteristics justify 100 percent detection. Most plants therefore run both: a measurement program that keeps the coater centered in its specification, and detection systems that catch the discrete events the measurement program cannot see.

Question 38

Where in a three-piece line is coating quality measured?

Accepted Answer

At four natural points: the coated sheet, the welded body before striping, the striped body after cure, and the finished can. Each answers a different question. On the coated sheet you measure film weight or thickness and check plain margin position and width, because everything downstream inherits those. After welding you assess weld appearance and the extent of bare metal the stripe will have to cover. After striping and cure, check stripe coverage and continuity. This is where 100 percent detection normally goes in, because a missing stripe is a product-contact failure. On the finished can you take enamel rating as the integrated result of everything that happened, plus dimensional and visual checks. Measuring only at the end tells you that something went wrong; measuring at all four tells you where.

Question 39

How often should a coating thickness gauge be calibrated?

Accepted Answer

There is no universal interval. Risk, usage and the evidence in your own records set it. The starting point is normally the manufacturer's recommendation and whatever your quality system or customer specification requires. From there, data adjusts the interval. If the as-found results at each calibration put the instrument comfortably inside its limits, you can extend the interval with justification. If instruments regularly come back out of limit, the interval is too long. You then have a product decision to make about everything measured since the previous calibration. Independently of the formal interval, daily or shift-start checks against a check standard are normal practice. They catch a damaged probe or a knocked instrument on the day it happens, not months later.

Question 40

Why do lab and line measurements of the same can disagree?

Accepted Answer

Because they are rarely measuring the same thing in the same condition. Before concluding that one instrument is wrong, check five things. Location: did both readings come from the same defined position on the can, described precisely enough that two people find the same spot? Condition: was the can clean, dry and at the same temperature in both cases, or did one come warm from the oven? Method: same settings, same substrate selection, same number of readings, same averaging rule. Reference: do both instruments trace back to the same standards, and when did each last record an as-found check? Sample: is it literally the same can, or two cans from the same batch — because can-to-can variation then explains the difference entirely. Most lab-versus-line disputes resolve at location or condition.

Question 41

Calibration, verification and adjustment — what is the difference?

Accepted Answer

Calibration compares your instrument against a traceable reference and records the difference. Verification confirms the instrument is still within its permitted error. Adjustment changes the instrument so that it reads correctly. People frequently treat them as one activity, and audits are where that catches up with them. The practical sequence runs in five steps. Calibrate against traceable standards. Record the as-found result. Adjust if the error falls outside the permitted limit, then record the as-left result. Finally, verify at intervals between calibrations using check standards. The as-found record is the valuable one. It tells you whether the product you measured since the last calibration got the right verdict.

Question 42

What should I check when Gauge R&R is poor?

Accepted Answer

First establish whether the study itself was valid, then split the problem into repeatability and reproducibility, because they have different fixes. Study design. Did the parts span the real process range, or did they all come from one pallet? A study run on near-identical cans makes the measurement system look bad because there is no part variation to compare it against. Were the parts, operators and trials properly randomised and the parts identified so operators could not recall previous readings? Did every reading come from genuinely the same location on each part? Poor repeatability points at the gauge and the presentation. Look for worn, contaminated or damaged probe or contact surfaces. Look for inadequate or inconsistent fixturing, so the part sits differently each time. Then check resolution against the tolerance, environmental stability, drift within the study, and surface condition varying between readings. Poor reproducibility points at method and training. Look for no written measurement procedure, or operators locating the measurement point differently. Look for different dwell, pressure or seating technique, different views on when a reading has settled, and different handling or cleaning of the sample before measurement. The fix is usually a specific, illustrated work instruction and re-training rather than a new gauge. Finally, check the tolerance the study runs against. A measurement system that suits a wide specification can be inadequate for a tight one. The same gauge can pass one study and fail another for that reason alone.

Question 43

How does eddy current measurement differ from capacitance measurement?

Accepted Answer

They interrogate different physics and therefore suit different substrates and coatings. Eddy current methods use a coil to induce currents in a non-ferrous conductive substrate such as aluminum. The coupling between coil and metal changes with distance, and that distance is the coating thickness. Magnetic induction methods do the equivalent job on ferrous substrates such as steel and tinplate. Capacitance methods measure the coating itself as a dielectric rather than measuring the distance to the metal. Four things drive the choice: the substrate, whether the coating conducts or carries pigment that interferes, the geometry of the part, and whether the measurement has to stay non-destructive. What all of them share is the need for a reference: a known substrate, a known zero and a set of traceable thickness standards.

Question 44

How does capacitance coating thickness measurement work?

Accepted Answer

A non-conductive coating on a conductive substrate behaves as the dielectric of a capacitor. A probe electrode on the coating surface forms one plate and the metal underneath forms the other. The coating between them sets the capacitance. Capacitance falls as the dielectric gets thicker, so the measured capacitance converts to a coating thickness. What follows from the principle is what matters in practice. The method needs a conductive substrate and an electrically connected reference, so it works on tinplate, TFS and aluminum but needs the can properly grounded. It responds to the dielectric properties of the coating, so a different coating chemistry generally needs its own setting or calibration. It averages over the area the probe covers, so it reports a mean film over that spot rather than the thinnest point within it. And it is sensitive to anything that changes the gap or the contact — surface contamination, moisture, an uneven or curved surface, probe tilt, or pressure. Measuring at defined positions with a clean, dry, correctly presented surface is not a formality; it is what makes the readings comparable.