Ask five slitting shops how many times a rotary knife can be reground and you will get five numbers, all of them confidently stated and none of them transferable. The heaviest gauge that knife is asked to cut sets the count, not the knife itself. Send two identical 20 mm knives out, one to a line running 4 mm hot rolled and one to a line running 0.6 mm cold rolled, and the second will outlive the first by years.

A more useful question, then. How much thickness can this knife afford to lose before it turns too flexible for my worst job? That one has an answer, and working it out takes about a minute.

This page sets out the arithmetic, the stock removal figures behind it, the inspection steps most shops skip when they sharpen in house, and the six conditions that genuinely mean scrap.

For straight guillotine blades rather than rotary knives, the decision runs on different rules and is covered in our guide on when to sharpen or replace shear blades.

What slitter knife sharpening actually means

On a coil slitting head there is no bevel doing the cutting. The working edge is the corner formed where the outside diameter runs into the flat side face, a nominal 90 degree corner all the way round the knife. Cutting happens as the top knife corner and the bottom knife corner pass one another with a set horizontal clearance and vertical overlap.

Wear turns up in three places at once:

1. The corner goes round. By the time it carries a radius of about 0.05 mm it has stopped shearing thin material cleanly and started tearing it.
2. A band appears on the OD land, polished or lightly scored, a few millimetres back from the edge, where the strip drags as it separates.
3. A matching band appears on the side face, in the same zone, put there by strip contact and side load.

Which is why sharpening a slitter knife has little in common with sharpening a bevelled tool. There is no edge angle to restore. You take material off until a fresh, square, undamaged corner exists again, and then you stop.

Flip the knife before you send it out

Every sharpening cycle spends one of two separate budgets, and knowing which one you are drawing down changes what you do next.

Countdown 1: thickness, spent by face grinding

Face grinding surfaces the flat side to bring a fresh corner up. It costs thickness. Thickness is what stops the knife bending under load, and a knife that bends puts width variation into the mult and camber into narrow strip. Every face grind walks the knife closer to the point where it is too soft for the gauge you run.

Countdown 2: diameter, spent by OD grinding

OD grinding takes the periphery in to clear chipping, ovality or a deep wear land. It costs diameter, and with it the amount of knife standing proud of the spacer.

Now the counter-intuitive bit. Taking OD off makes a knife stiffer, not weaker, because there is less unsupported overhang past the spacer. Jim Wilcox made the same observation in The Fabricator: veteran setup men will happily tell you the line cambered worse on new tooling than it does now, years of grinding later. He is equally clear on why OD grinding stays a last resort, since the rubbers have to come in along with the steel and you are burning knife life to do it.

So the two countdowns pull opposite ways on stiffness, and whichever one empties first retires the knife from that particular job.

Working out your own regrind budget

The accepted sizing rule for new knives puts thickness at three to five multiples of the heaviest gauge the knife will ever see, on the logic that tooling has to survive its worst day rather than its average one. At the bottom of that band, 3x holds up on light work and starts flexing at the top of the gauge range. At the top, 5x is comfortable, and to most purchasing managers it looks like money wasted.

Which is where the buying decision usually goes wrong.

Turning budget into a regrind count means dividing by the stock you take off each visit. A full cycle grinds both faces, so it costs roughly double the per-face figure.

Look hard at that bottom row. Running a knife until it chips writes off about three quarters of the tooling life you paid for. Chipped knives are recoverable. They are simply four to five times more expensive to recover, because the damage sits deeper than the wear does, and no other habit on a slitting line costs so much for so little gain.

It also explains why the three-to-five figure quoted around the industry is neither wrong nor much use on its own. Thin knives on light gauge really do land there. A 25 mm knife on heavy hot rolled, pulled at the right time, will take sharpening for a decade. Same rule, different gauge, wildly different answer.

How much to take off, and how little is too little

One principle governs the whole operation: take off what the damage demands and not a micron more. DIENES gives its regrinding customers the same instruction, for the obvious reason that every extra micron is life you do not get back.

There is a floor as well as a ceiling, and the floor catches more people. Grind too shallow and you leave the fatigued layer sitting directly under the new corner, so the knife goes back on the arbor with fractured metal a few hundredths below its cutting edge. It will chip inside a shift. Converting-industry guidance puts that damaged depth somewhere around 0.25 to 0.38 mm. On coil knives the depth tracks how hard the tooling was worked rather than any fixed figure, which is exactly why you judge the ground face instead of grinding to a number.

The test is visual and takes five seconds. Turn the knife under a light. The ground face should read uniform right around the circumference, with no ghost of the old wear band anywhere on it. A faint ring still showing means you stopped short.

Slitter blade sharpening in sets, never one at a time

Where several knives share an arbor, they go to the grinder together and come back at one common diameter. DIENES is blunt about this, and the reasoning is simple. Overlap has to sit even across the full width of the head, which it will not do if one knife returns a different size from its neighbours.

Pull a single knife, take 0.4 mm off its OD, drop it back into an otherwise matched stack, and one cut on that head is now running a different penetration from every other cut. The symptom is an edge quality difference on two mults and nothing anywhere else, which is the kind of fault that gets blamed on the coil supplier for three weeks running.

Thickness needs the same discipline for a different reason. Knife thickness feeds the stack arithmetic that sets slit width and horizontal clearance, so once a knife has been sharpened:

• Measure each knife on its own to 0.001 mm and write down what you actually got, not what the stamp says.
• Mark the knife with its measured thickness, or at minimum with a set identifier.
• Rebuild the knife and spacer stack off measured values.
• Re-check horizontal clearance and vertical overlap for the material about to run, because thinner knives shift the gap between the top and bottom shearing planes.

Nine times out of ten, a complaint that the regrind came back bad turns out to be a setup nobody recalculated.

Grinding burn, the failure nobody inspects for

A slitter knife at 60 to 62 HRC has been vacuum hardened and tempered into a specific structure. Grinding dumps heat into a surface layer a few thousandths of a millimetre deep. Two things can go wrong in that layer and a micrometer will show you neither.

Over-tempering. The surface climbs past its tempering temperature and softens, leaving a low-hardness skin on the one face that forms the cutting corner. The knife looks sharp, feels sharp, and fails early. DIENES carries the same warning on its own service page: heat the knife while resharpening and you lose hardness at the edge.

Rehardening burn. More heat again, then coolant quenching it back down, and you get untempered martensite carrying high tensile residual stress. That is where grinding cracks come from, on a hardened disc that then goes back onto an arbor and turns at line speed.

Controlling it comes down to four habits:

• Flood coolant, every pass. Dry grinding a hardened tool steel face is a scrap decision taken early.
• Soft, open, friable wheel. Hard wheels glaze, stop cutting and start rubbing, and rubbing is heat. Dress often.
• Light infeed, then spark out. Finish on light passes rather than one heavy one.
• Alternate the faces. Grinding one side heavily before the other loads the knife with lopsided residual stress and pulls it into dish.

Detecting it takes four checks, in ascending order of effort:

ISO 14104:2017 is written around gears, but the chemical temper etch it specifies is the reference method for finding localised overheating on any ground hardened steel surface, and the standard notes it reads changes in surface hardness that ordinary hardness testing will miss.

The two finishing steps that get skipped

Take the burr off. Grinding raises a burr and a wire edge along the corner. Send the knife back to the arbor still carrying it and, as DIENES points out, the edge starts shedding microscopic particles almost straight away. You lose the corner you just paid for and the debris goes into the strip. Hone it.

Demagnetise. Face grinding happens on a magnetic chuck, and a knife that leaves it still magnetised will pick up steel fines on both faces for the rest of its working life. Those fines end up between knife and spacer, where they act as a stack error and a runout source that no amount of careful measuring will ever find. Demagnetise before the knife goes near an arbor.

Knives that were lapped when new need lapping again after sharpening, not just grinding and shipping. Ground-only faces carry a coarser, directional finish that changes how the strip releases and how the slit edge burrs. Our note on grinding against lapping covers what that difference does to cut quality.

When to scrap the knife

Six conditions. The first is arithmetic, the other five are pass or fail.

1. Thickness has dropped under 3x the heaviest gauge that knife will run, or under the line builder’s stated minimum, whichever is the larger number.
2. Any crack, anywhere. A cracked hardened disc turning at line speed under thousands of kilograms of separating force stops being a tooling question. Bin it.
3. Hardness has fallen in the ground zone and clearing it would breach the thickness floor. If sound metal sits further down than your remaining budget reaches, the knife is done.
4. Bore gone oval or keyway peened. That is a runout source no amount of face grinding will touch. Re-boring and bushing occasionally pays on large heavy knives and almost never pays on standard sizes.
5. Permanent dish that will not come flat inside the budget left. Side load cones a knife over time. Measure parallelism and flatness before quoting the sharpening, not after.
6. OD too small to reach the penetration your line needs with the spacers and stripper rings you actually own. Modern heads run around 38 mm of knife proud of the spacer, older ones 44 to 51 mm. Fall below that and the stripper rings have to come down as well, at which point you are grinding rubber to save steel.

Do not scrap it, demote it

Condition 1 belongs to the job, not to the knife, so most tooling that ends up in the scrap bin is only finished for one job.

Take a 20 mm knife that has come down to 11 mm. On 4 mm hot rolled it is out, since the floor there is 12 mm. On 2 mm cold rolled the floor is 6 mm, so it still holds 5 mm of budget, worth another 20 to 30 sharpening cycles. That is real money sitting in a skip.

Build the second tier on purpose:

• Segregate by measured thickness, not by how old or how battered the knives look.
• Push thinner knives towards lighter gauge as their budget shrinks.
• Push smaller-OD knives towards narrow mults, where the shorter overhang works for you. Wilcox recommends the same split: older, smaller-diameter knives on narrow work, new knives on wide material where camber is much harder to induce.
• Keep one deliberately undersized knife. A scrap knife taken at least 12 mm under working OD, parked outboard of the trim cut, is a recognised fix for trim knives deflecting outward with nothing pushing back.

Sharpening in house or sending them out

Cost per knife is the wrong test. The real one is whether you can tell a good sharpening job from a bad one after the knife comes back.

In house works if you have coolant discipline, a dressed wheel, an operator who alternates faces, and the metrology to check flatness, parallelism and thickness afterwards. Turnaround drops from weeks to hours and set matching stays under your control.

Sending out works when you cannot verify the result. An over-tempered knife is indistinguishable from a good one until it chips out three days into a run, and by then the downtime and scrapped coil have cost more than years of grinding bills. Same logic if the tooling was lapped originally and you have no lapping capability, because grinding it back without restoring the finish downgrades the knife a little on every cycle.

A split arrangement suits most shops: routine face grinding in house, and everything else out, meaning OD work, set re-matching, and any job that has to come back with hardness or flatness certification.

Slitter knife sharpening at Maxwell

Worn rotary knives come into our Rajpura plant from lines all over the world, and we take Maxwell and non-Maxwell tooling alike. Each knife goes back to its drawing dimensions, gets its corner geometry and land width restored, gets lapped, gets its hardness re-verified, and goes onto the CMM before it is packed. An inspection report travels with it. Ten working days from the day the box lands, at roughly 40 percent of what the same knives would cost new.

Send a used sample or a drawing and we will tell you plainly which knives still have budget in them and which have run out. Where the answer is new tooling, our rotary slitter blades and knives are made under one roof in D2, H11, HSS M2, M35 and PM grades, vacuum hardened, held to ±0.0015 mm on thickness.

Email [email protected] or call +91 82646 21011. Technical queries get an answer inside one working day.