A slitting line that starts throwing burrs is rarely a mystery for long, but it is almost always misdiagnosed on the first attempt. The operator sees a rough edge, assumes the knife is worn, fits a new set, and the burr comes back within a shift. Three knives later, the real cause is still sitting in the machine: a holder with 0.03 mm of radial play, a spacer stack that was never cleaned, or an overlap figure that was copied from a paper job onto a 12 micron PET film.
This guide is written for the person who has to stop the line and find the answer. It separates the six blade-side causes from the four process-side causes, gives the measurement method for each, and sets out a diagnostic order that finds the fault in one pass instead of five. Every number here is a starting point to be verified against your machine builder’s specification and your own edge quality target, not a universal constant.
One clarification before anything else, because it causes more wasted time than any other confusion in slitting: a burr is not the same defect as a turned edge, a nick, or angel hair. They look similar in a photograph and they have completely different causes. Fixing the wrong one is how a two-hour problem becomes a two-week problem.
Four Edge Defects That Get Called “Burrs”
Before adjusting anything, classify what is actually on the edge. A 20x loupe or a USB microscope is enough, and it takes less time than changing a knife.
| Defect | What it looks like | Usual cause | First thing to check |
|---|---|---|---|
| Burr | A raised lip of material standing proud of the cut face, on one side or both | Shear is incomplete; the web is being bent and broken instead of cut | Overlap and edge sharpness |
| Rolled or turned edge | The edge is folded over, often with a thin flag of material lying flat against the face | Side clearance too open, or side load too low to hold the knives in contact | Side load and holder bearing play |
| Nick or chip mark | A short, repeating notch at a fixed spacing down the web | A damaged spot on the knife edge, or a hard particle passing through | Spacing between marks against knife circumference |
| Angel hair or fuzz | Fine filaments, hairy edge, or fibre pull-out on film and nonwoven | Edge radius too large for the web; tearing rather than shearing | Edge radius under magnification, then overlap |
The spacing of a repeating defect is the single most useful piece of evidence on the floor. If the marks repeat once per knife revolution, the fault is on the knife or its mounting. If they repeat once per unwind roll, the fault is in the incoming material. If they appear at random, it is process instability, usually tension or web guiding.
The same classification applies outside film and foil. On a corrugated line, what operators call a burr is usually a crushed edge at the slot, produced by a carton slotter knife that has lost its edge or is running out of alignment with the anvil; the cut looks ragged because the flute has been compressed rather than severed. Our note on corrugated blades covers the difference between a dull cut and a worn slot, which are often confused in exactly the same way.
The Six Blade-Side Causes
1. The edge is duller than it looks
An edge that still feels sharp to a thumb can be well past the point where it shears cleanly. What matters is the edge radius, not the visual impression of sharpness. As the radius grows, the contact stress at the nip falls below the level needed to initiate a controlled crack, and the material starts to deform and tear instead of separating. On thin film the result is angel hair; on coated board it is a dusty, feathered edge; on aluminium foil it is a raised lip that is sharp enough to cut a glove.
Dull-edge burrs have a signature: they appear on both sides of the cut, they get progressively worse across a run rather than appearing suddenly, and they disappear completely after a knife change. If burr complaints climb toward the end of a shift, that pattern is the diagnosis. The fix is a resharpening interval set by linear metres cut or by hours in the cut, not by an operator’s judgement at the machine.
2. Overlap is set for the wrong material
Overlap, also called penetration or depth of cut, is how far the top knife passes beyond the cutting plane of the bottom knife. It is the parameter most often copied from one job to the next, and it is the one that punishes copying hardest. A setting that gives a clean edge on 300 gsm board will crush a 12 micron film and generate heat, while an overlap thin enough for film will leave uncut fibres on board.
One-sided burrs are almost always an overlap problem rather than a sharpness problem. Too much overlap pushes material ahead of the edge instead of shearing it, producing a thickened, work-hardened edge. Too little leaves a gap where the web tears as the two faces separate. The correct value is the smallest overlap that produces a complete, clean cut at production speed, and it should be re-verified after any machine rebuild, because a gauge that drifts by 0.03 mm is enough to change the result.
3. Side clearance has crept open
In true shear slitting the two knife faces should run in light contact, held there by side load rather than by a gap you can measure with a feeler gauge. A measurable gap lets the web fold into the space instead of being sheared, which produces the classic turned or wavy edge. Clearance opens gradually as holder bearings wear and as knife shafts pick up runout, so the symptom tracks speed rather than time: burrs get worse as the line runs faster, and improve when it is slowed down.
This is the cause that survives knife changes, because a new knife in a worn holder still has a gap. If a customer reports that the problem came back after fitting new blades, bearing play and shaft runout are the first two items to measure, not the third.
4. Runout you cannot see
Total indicated runout on a mounted knife should stay small enough that the contact point does not move in and out of the bottom knife face during a revolution. The practical ceiling quoted by holder and knife makers is around 0.01 mm on the mounted knife, and holder radial play above roughly 0.02 mm is enough to let the knife wander under cutting load. Neither figure is visible to the eye, and neither is fixed by a new knife.
The tell-tale is a burr pattern that repeats at the knife circumference: one bad patch per revolution, evenly spaced down the web. Check runout on the knife OD, on the mounting face, and on the shaft itself. A knife that sits 0.02 mm off on a dirty mounting face behaves exactly like a bent knife.
5. The blade material is fighting the substrate
This cause does not look like wear, which is why it is missed. A harder blade on an abrasive filled film holds its edge longer, but the same hard blade on a sticky, adhesive-backed or highly filled substrate can chip at the edge instead of wearing evenly, leaving micro-burrs that appear random. One roll is clean, the next is not, and nothing on the machine has changed.
The match is between edge toughness and substrate, not simply between hardness and abrasion. Where wear is the dominant failure, a harder grade or a coating carries the load; where chipping is dominant, extra hardness makes it worse. Our material comparison for industrial blades sets out that trade-off, and the coating overview explains when a coated edge helps and when it merely moves the failure mode.
6. Contaminated mounting faces and spacer stacks
A speck of dried adhesive or a burr on a spacer face is 5 to 20 microns of error that nobody measures. In a gang set with fifteen knives, those errors accumulate across the stack, so the outer stations run at a different effective overlap from the centre. This is why burrs on wide lines are often worse at the edges of the web than in the middle, and why the problem disappears after a full strip-down and clean.
The rule is that every contact face in the stack is a precision surface. Wipe, inspect, and store knives on a flat protected surface. A stack assembled from a parts bin will hold a tolerance only by luck.
| Cause | How to confirm it | What you will see |
|---|---|---|
| Dull edge | Inspect edge under 20x magnification; measure edge radius | Burrs on both sides, worsening through the run |
| Wrong overlap | Measure penetration with a setting gauge; compare to material spec | One-sided burr, or crushed and shiny edge |
| Open side clearance | Feeler gauge at the knife faces; check side load reading | Turned edge, worse at higher speed |
| Runout | Dial indicator on knife OD, mounting face and shaft | Repeating defect once per revolution |
| Material mismatch | Compare failure mode: even wear versus chipping | Intermittent burrs, no machine change |
| Contaminated stack | Strip and inspect spacer and knife faces | Burrs worse at outer stations of a wide line |
The Four Process-Side Causes
If the knives and the stack check out, the fault is upstream or downstream of them. These four account for most of what remains.
Web tension. Too little tension lets the web flutter as it enters the slitting station, so the material meets the knives at a varying angle. Too much tension causes neck-in: the web narrows before it reaches the cut, and the edge is deformed by the time it is severed. The signature of a tension problem is a burr that varies across the width of the roll or appears only at certain roll diameters.
Web guiding. A misaligned web approaches the knife at an angle, which changes the effective shear geometry on one side of the line. On narrow strips this shows up as one good edge and one bad edge from the same pair of knives.
Speed. Above the point where the web is stable, the knives generate heat faster than the web can carry it away. Heat softens some films and embrittles others, and both outcomes produce edge defects that vanish at lower speed. This is the cheapest test available: run the same material at 70 percent speed and see whether the burr goes away. The same thermal argument applies to the cross-cut knives on packaging lines, where a dull cut-off blade crushes the seal area and the defect is usually blamed on the sealing jaw instead.
Incoming material. A new roll with a different coating weight, a harder calender density, or a slightly different foil temper will produce different edge quality on identical settings. Compare the certificate of analysis before blaming the knives.
| Process cause | Symptom pattern | Quick test |
|---|---|---|
| Low tension | Burr varies with roll diameter; web flutter visible before the cut | Raise tension in small steps and watch the entry |
| Excessive tension | Finished slit width narrower than set; hard raised edge | Measure slit width against knife spacing |
| Web guiding error | One edge good, one bad from the same knife pair | Check edge position sensor and entry angle |
| Speed too high | Burr appears only at production speed | Run at reduced speed for one roll |
| Material variation | Problem starts with a new roll, not with a knife change | Compare COA between old and new roll |
Diagnose in the Right Order
The sequence matters more than the individual checks, because changing two variables at once destroys the evidence. Work down this table and stop at the first line that explains what you are seeing. Each step is a measurement, not an opinion.
| Order | Check | Acceptance | If it fails |
|---|---|---|---|
| 1 | Classify the defect type under magnification | Defect identified as burr, turn-over, nick or fuzz | Repeat with better lighting before proceeding |
| 2 | Measure defect repeat spacing | Matches knife circumference, roll circumference or random | Use the spacing to select the branch below |
| 3 | Inspect knife edge for radius, nicks, heat marks | Edge uniform, no visible chipping | Change or resharpen the knife |
| 4 | Measure mounted runout on knife OD and face | Within builder’s limit, typically around 0.01 mm | Clean mounting face, re-seat, or replace holder |
| 5 | Check holder radial play | Below roughly 0.02 mm | Rebuild or replace the holder |
| 6 | Verify overlap against material specification | Within the window for that web | Reset in small increments, one variable at a time |
| 7 | Verify side load and contact between faces | Faces in light contact, no measurable gap | Increase side load to the minimum that stabilises the cut |
| 8 | Check tension, guiding and speed at the entry | Stable web, entry angle square to the knife | Correct the process parameter, not the knife |
| 9 | Compare incoming material to the previous roll | Same grade, coating weight and temper | Quarantine the roll and re-test |
The overlap value used in step six is material specific, and the working windows by substrate thickness are set out in our clearance and overlap chart.
Two rules make this sequence work. First, write down the settings before you change anything, because a line that “ran fine last week” is worthless evidence if nobody recorded what it ran at. Second, change one variable per trial roll. The temptation to adjust overlap and side load together is strong and it is the reason many plants never find the root cause.
How to Measure a Burr Properly
“It looks rough” is not a specification, and a supplier cannot act on it. Burr height needs a number and a method, because the same edge measures differently depending on how it is checked.
| Method | What it measures | Resolution | Best use |
|---|---|---|---|
| Optical microscope with scale | Burr height in the plane of the cut | A few microns | Routine line-side check and defect classification |
| Profile projector | Edge silhouette and burr profile | 1 to 2 microns | Batch consistency and supplier comparison |
| Cross-section microscopy | Sub-surface damage, rollover, delamination | Destructive, high detail | Failure analysis and customer complaints |
| Laser or confocal profilometer | 3D edge topography | Sub-micron | Battery, electronics and medical work |
| Tape pull test | Loose particles and fibre release | Qualitative | Quick screening on coated and dusty webs |
Whatever method you use, fix the sampling position. Burr height varies along the slit length and across the width of the roll, so a single reading at the start of a roll proves nothing. Take readings at the start, middle and end of the run, on both edges of at least three strips, and record the maximum rather than the average, because the maximum is what punctures a separator or fails a customer’s incoming inspection.
What Counts as Acceptable
There is no universal burr limit, and any supplier who quotes one without asking about your product is guessing. What exists is a set of industry conventions, and they differ by an order of magnitude between sectors.
| Sector | Common acceptance reference | What drives the limit |
|---|---|---|
| Flexible packaging film | No visible raised edge at 20x; no angel hair | Print quality, seal integrity, web breaks |
| Paper and board | Edge free of dust and fibre pull-out | Converting dust, printing blanket damage |
| Aluminium foil | No lip that catches on a glove or foil surface | Pinholes, lacquer coverage, operator safety |
| Nonwoven and hygiene | No fibre protrusion; clean severance | Consumer feel, leakage, lint |
| Lithium-ion electrodes | Burr height below roughly half the separator thickness; production targets in the 3 to 16 micron range depending on cell design | Separator puncture and internal short circuit |
| Printed circuit board | Edge burr small enough not to affect routing and plating | Delamination, plating defects |
| Corrugated board | No crushed edge; score line clean through the flute | Box compression strength, print registration |
The battery figure deserves a note, because it is the one place where the specification is genuinely tight and where suppliers frequently overstate certainty. Published references vary: some cell makers work to a fraction of separator thickness, some quote a fixed micron figure for a specific platform, and some apply a fraction of electrode thickness. All three are legitimate, and they are not interchangeable. If your product is a cell, the acceptance limit is set by your separator, your assembly pressure and your safety margin, and it should be written into the drawing rather than inherited from a supplier’s brochure.
Sectors with a hygiene or regulatory dimension add a second layer. A blade on food-contact equipment has to be geometrically clean as well as sharp, because a burr that traps product residue is a sanitation problem rather than only a quality one. That is why food and hygiene lines often specify a polished edge and a documented material grade even when the measured burr height would pass a dimensional check.
Why Burrs Grow as the Tool Wears
Understanding the mechanism explains why some fixes work and others only move the problem. A clean slit requires the edge to concentrate stress until the material separates by controlled crack propagation. That works while the edge is sharp relative to the material thickness. As the edge rounds, the contact area grows, stress falls, and the material starts to deform plastically before it separates. Plastic deformation at the edge is a burr.
Two consequences follow. First, the same edge condition that cuts 200 micron board cleanly will produce a burr on 12 micron film, because the ratio of edge radius to thickness is what matters, not the radius alone. This is why one line’s “still good” knife is another line’s scrap. Second, burr growth is not linear: there is a threshold where edge quality falls away quickly, and running a knife past that threshold damages the mating knife as well. A knife that is changed at the first sign of rounding costs less than one that is run until the edge fails. Setting a sensible interval is the subject of our service life guide.
There is a third consequence that is often missed. Burr height is also a function of the stability of the cut, not only of sharpness. A perfectly sharp knife in a holder with play, or on a line with unstable tension, produces a variable burr because the effective overlap changes during the revolution. This is why measuring runout and bearing play is not an optional refinement: without those two numbers, edge quality data cannot be interpreted.
The Eight-Step Recovery
When a line is producing out-of-spec edges, this is the sequence that recovers it with the fewest wasted rolls.
- Stop and document. Photograph the edge, record the settings, note the roll number and the knife set. Without this you cannot prove the fix worked.
- Classify the defect. Burr, turn-over, nick or fuzz. Everything downstream depends on getting this right.
- Measure the repeat spacing. Knife circumference, roll circumference or random.
- Inspect the knife. Magnification, edge radius, chips, heat discolouration, mounting face condition.
- Check the geometry. Runout, holder play, overlap, side load. Measure, do not adjust by feel.
- Run a trial at reduced speed. If the defect disappears, you have a heat or stability problem, not a sharpness problem.
- Change one variable. Overlap first if the edge is one-sided, side load first if it is turned over, knife first if the edge is rounded. One roll per change.
- Lock the result. Write the winning settings onto the changeover sheet and log the knife hours. A fix that lives only in an operator’s memory is not a fix.
| Step | Tool needed | Evidence it produced |
|---|---|---|
| Document | Phone camera, settings sheet | Baseline for comparison |
| Classify | 20x loupe or USB microscope | Defect type |
| Repeat spacing | Tape measure, knife drawing | Knife, roll or random |
| Knife inspection | Microscope, edge radius gauge | Edge condition |
| Geometry | Dial indicator, feeler gauge, setting gauge | Runout, play, overlap numbers |
| Speed trial | Line controls | Thermal or stability signature |
| Single change | Trial roll | Cause confirmed or excluded |
| Lock | Changeover sheet, knife log | Repeatability across shifts |
When to Change the Knife Instead of Adjusting
Adjustment cannot recover a worn edge, and continuing to adjust wastes production time. Pull the knife when any of the following is true.
| Criterion | What you are looking for | Why it matters |
|---|---|---|
| Edge radius | Visible rounding under 20x magnification | Shear has become deformation |
| Chipping | Any nick visible on the cutting edge | One nick propagates a defect down the whole roll |
| Heat marks | Blue or straw discolouration on the body | Indicates rubbing and loss of temper at the edge |
| Diameter wear | Reduction beyond the range where the holder can still set correct overlap | Geometry can no longer be achieved |
| Mating knife condition | Wear on the bottom knife face | A dull top knife damages the bottom knife |
| Burr trend | Edge quality falls off before the planned interval | The interval is too long for this material |
For lines running abrasive filled film, white or metallic inks, or mineral-filled compounds, the interval should be set from the burr trend rather than from a fixed calendar. Track burr height against metres cut for three knife sets and the curve will tell you where the threshold sits. Our guide to extending industrial blade service life covers the same logic for other blade types, and the heat treatment and hardness guide explains why two knives at the same nominal hardness can wear at different rates.
Matched sets are worth the extra line on the purchase order. A new top knife running against a worn bottom ring has no stable reference face to shear against, so the setting drifts within a shift and the burr returns. Our circular slitter blade range is quoted in sets for this reason, and the round blade category lists the matching ring options alongside each knife.
Prevent It: The Changeover Sheet
Repeatability across shifts is what protects output, and it comes from a written record rather than from experience. A changeover sheet that captures the last known good settings turns every future setup into a verification instead of an experiment.
| Field | Why it is recorded |
|---|---|
| Material grade, thickness and supplier batch | Separates material variation from tooling variation |
| Knife set ID and hours run | Links edge quality to knife life |
| Overlap setting and how it was measured | Prevents a “feel” setting from entering the process |
| Side load reading | Detects holder drift over time |
| Runout at each station | Catches shaft and mounting damage early |
| Line speed and tension | Needed to reproduce a good run |
| Burr height at start, middle and end | Shows whether the defect is time-dependent |
| Who set it and when | Makes the record actionable |
What a Burr Actually Costs
Burrs are usually treated as a quality annoyance until someone adds up the cost. Four items dominate, and only the first is visible on a scrap report.
| Cost item | How it appears | Typical scale |
|---|---|---|
| Scrap and downgrade | Rolls rejected at final inspection or by the customer | Visible, but often under-recorded |
| Downtime | Stops for knife changes and trial rolls | Grows quickly when diagnosis is by substitution |
| Knife consumption | Premature changes and regrinds | Doubles when the real cause is a worn holder |
| Dust and contamination | Cleaning, filter changes, print defects downstream | Rarely attributed to slitting at all |
The most expensive pattern is not a bad knife. It is a good knife in a worn holder, replaced five times before anyone measures bearing play. That is a 30-minute measurement standing in for a month of knife spend.
Tool cost also moves with the carbide market. Where a line runs tungsten carbide knives, the cost per thousand metres depends on the price of the insert at the time of purchase, and tungsten carbide price movements are worth tracking before a knife interval is renegotiated. A buying decision based on last year’s price per piece is usually the wrong decision this year.
What to Send a Blade Supplier
Most edge problems that reach a supplier are unanswerable because the enquiry contains only a material name and a complaint. Eight parameters decide whether a knife recommendation is meaningful.
| Parameter | Why the supplier needs it |
|---|---|
| Substrate and exact thickness | Sets the overlap window and the edge geometry |
| Coating or filler content | Determines abrasiveness and the wear mechanism |
| Line speed | Affects heat generation and holder rating |
| Slitting mode: shear, crush or score | Different geometry and different tolerances |
| Knife diameter, thickness and bore | Must match the holder and the arbor |
| Current overlap and side load settings | Shows whether the problem is geometry or edge |
| Target burr height or edge standard | Defines what “good” means for your product |
| Photo of the edge at magnification | Classifies the defect in one look |
With those eight items, a supplier can usually identify the failure mode before quoting. Without them, any recommendation is a guess dressed as advice. The supplier side of the same decision is covered in our guide to qualifying an industrial blade supplier, which lists the audit questions worth asking before a first order. Our slitting knife selection guide works through the geometry side of the same decision, and the blade troubleshooting guide covers failures that are not edge-related.
Matching the Knife to the Line
The knife itself has to suit the mode of cutting. Shear slitting uses two rotary knives in a scissor action and is the standard for film, foil, paper and nonwoven. Crush cutting uses a knife against a hardened anvil or a second shaft and is common on thinner, softer webs where a shear setup is hard to hold. Score cutting is a pressure process and depends on edge sharpness far more than on geometry. Each has a different burr mechanism, which is why a fix that works on one line can fail completely on another that looks similar.
For shear lines the working parts are the circular slitting knife on top and a matched bottom ring below it. The pair has to be treated as a set: a new top knife against a worn bottom ring will not hold an edge for long. Where the application is abrasive, a tungsten carbide slitter blade changes the wear rate, and where the bottom ring has worn or been damaged, a wear-resistant slitting lower blade restores the reference face the top knife needs. Wide lines with many stations often use large round slitting blades to keep deflection under control at the outer positions. The full range sits under the round blade category, and buyers comparing grades can start from the tungsten and steel blade category.
Dekeer supplies shear and crush slitting knives in matched sets, with the bore, thickness and edge geometry held to the drawing rather than to a catalogue number. Because the same knife geometry behaves differently across substrates, we quote against the substrate and the target edge standard, not against a model name. If a line is fighting burrs, send the eight parameters above and the edge photo; that is usually enough to tell whether the answer is a different edge, a different grade, or a holder that needs rebuilding. The contact page reaches the technical team directly, and the product overview shows the full blade range by application.
Questions From the Floor
These are the questions that come up most often when a line is being diagnosed, with the short answer for each.
| Question | Answer |
|---|---|
| Why did the burr appear after a knife change? | The new knife is probably not seated correctly, or the holder has play that the old knife masked. Check mounting face cleanliness and runout before changing anything else. |
| Why does the burr get worse as speed increases? | Heat and dynamic instability. Side clearance that opens under load, or holder play, shows up at speed rather than at idle. Run a slow trial to confirm. |
| Can I fix a burr by increasing overlap? | Only if the cause is insufficient penetration. If the edge is being crushed or the line is overheating, more overlap makes it worse. |
| Is a burr on one side always an overlap problem? | Usually, but web guiding and holder misalignment produce the same one-sided pattern. Measure the entry angle before adjusting overlap. |
| How often should knives be resharpened? | Set the interval from metres cut and burr trend, not from a calendar. Start with the machine builder’s figure and shorten it if edge quality falls off first. |
| Do coated knives reduce burrs? | A coating reduces wear rate, which keeps the edge in the good zone for longer. It does not correct a geometry fault, and on some substrates it introduces chipping. |
| Can a burr be removed after slitting? | Not without changing the edge. Deburring a slit roll is not practical at production speed, which is why the cut has to be right the first time. |
| What is the most common single cause? | In our experience, a worn holder or contaminated mounting face rather than a worn knife. It is also the cheapest thing to measure. |
Where to Read Next
Three adjacent problems share the same physics. A doctor blade metering ink onto a gravure cylinder fails in the same two ways, a rounded edge and the wrong contact angle, and the diagnostic sequence is almost identical. Where the substrate is sensitive enough that even a light burr matters, ceramic blades are worth evaluating for edge stability rather than hardness. And where the line is a packaging machine rather than a slitter, a heat sealing cutting blade has its own edge requirements because the cut and the seal happen in the same stroke. Buyers who need a non-standard geometry usually start from a custom wear-resistant blade drawing rather than from a catalogue size.
Where the Hardware Comes From
Nothing in this guide replaces a matched knife set in a sound holder. The shear hardware is listed under the round blade category, abrasive applications are covered in the tungsten and steel range, and a line that needs a non-standard geometry starts from a drawing rather than a size. Technical questions with an edge photo attached reach the team fastest through the contact page.
The Short Version
Classify the defect before touching the machine. A burr, a turned edge, a nick and angel hair have different causes, and only one of them is usually about sharpness. Measure runout and holder play before replacing knives, because a good knife in a worn holder produces exactly the defect you are trying to remove. Set overlap from the substrate, not from the last job, and change one variable per trial roll. Then write the winning settings down, because the only fix that survives a shift change is the one on the changeover sheet.