Symptom to Cause: A Bench Diagnosis Table for Knives That Stopped Cutting Well
Three-quarters of edges that will not hold were never properly deburred, or were sharpened at an angle that drifted mid-stroke. The rest are steel too soft for the angle, or a blade that is simply too thick behind a perfectly good edge. Each symptom below has a test that tells you which, and only one of them is fixed by sharpening more.
By the Knives Wow Editorial Team
If your knife is dull three days after sharpening, the overwhelmingly likely cause is a burr you never removed. It felt shaving-sharp on the bench because a wire edge is genuinely thin, and it collapsed on the first board contact because a wire edge is a flap of steel attached to nothing.
The second most likely cause is an angle that moved. Sharpening at 15° on the way down the stroke and 25° on the way back does not average out to 20°; it produces a rounded apex that never had a proper edge on it at all.
Beyond those two, there are only a handful of real failure modes, and they need different fixes. Rolled and chipped edges look similar and are caused by opposite things. And there is a whole category of problem where the edge is genuinely fine and the blade behind it is too fat — that one cannot be sharpened away, and grinding more will make it worse.
Four tests to run before you touch a stone
- Thumbnail catch. Draw a thumbnail lightly along the edge at right angles, heel to tip. A good edge bites evenly all the way. A dead spot means the apex is rounded there. A snag that comes and goes means a burr is still attached.
- Paper push-cut. Hold a sheet of printer paper vertically and push the edge straight into it with no drawing motion. Clean parting means apexed. Tearing or skidding means not.
- Light on the apex. Hold the edge up to a bright light and sight down it. A sharp apex has no width, so it reflects nothing. Any bright thread of light running along the edge is a flat, and the thicker the thread the more steel you have to remove.
- The tomato scale test. Rest the edge on a tomato sitting on a digital scale and press straight down. Under 200 g is sharp. Around 500 g is due a stone. Over 900 g is genuinely hazardous, because the knife will skid before it bites.
Run all four and write the results down. A knife that fails the light test but passes the thumbnail catch is telling you something completely different from a knife that does the reverse, and the table below only works if you know which is which.
One more measurement worth having in hand before you start: spine thickness 5 mm up from the edge, taken with calipers. That single number separates an edge problem from a geometry problem, and it is the number nobody thinks to take.
The diagnosis table
Symptom Diagnostic test Reading that confirms it Primary cause Secondary cause Fix, with grit and angle Recoverable? Dull within days of sharpening Thumbnail catch along the full edge, then paper push-cut Snag points that appear and vanish; paper tears rather than parts Burr never removed — a wire edge collapsed on first use Angle drifted, leaving a partially apexed edge Return to 1000 grit, re-raise a full burr, then deburr with descending strokes: 10, 5, 3, 1, 1, finishing edge-trailing on newspaper At home, 10 minutes Rolled edge — a bright line along the apex that deflects Sight down the edge under a bright light; run a fingernail across it A visible bright thread that moves as you tilt the blade; nail catches on one side only Bevel angle too shallow for the hardness of the steel Lateral load — scraping the board with the edge, or levering Two or three light alternating passes on a honing rod at the sharpening angle. If that fails, 1000 grit and add 2° per side At home, 2 minutes Chipped edge — visible nicks Phone camera on macro along the apex; measure the deepest nick Nicks 0.1 to 1.0 mm deep with clean square shoulders, not folded metal Hard contact: bone, frozen food, a stone in produce, or a ceramic board Angle too acute for the steel's hardness Chips under 0.3 mm: 1000 grit until gone. Over 0.3 mm: 400 grit at target angle plus 2°, then full progression At home under 1 mm; grinder above Dull immediately after sharpening Marker on the bevel, three strokes, inspect where ink survives Ink still present right at the apex, removed on the shoulder Angle too shallow — you polished the shoulder and never reached the apex Dished stone lifting the apex clear of the abrasive Flatten the stone. Raise the spine by 2 to 3 mm and work at 1000 grit until a burr appears heel to tip At home, 15 minutes Blade wedges and splits dense produce Calipers on the spine 5 mm above the edge; tomato scale test for comparison Over 0.7 mm at 5 mm up while the tomato test reads under 250 g Geometry, not edge. The blade is too thick behind a perfectly sharp apex Repeated sharpening over years has walked the apex up into thicker steel Thin the whole primary bevel on a 400 grit stone at roughly half the edge angle, then re-sharpen normally. Budget 40 minutes At home with patience; a grinder is faster and riskier Food sticks to the face of the blade Cut ten slices and count how many cling past the second slice More than six clinging Flat, wide blade face with high surface contact and a high Wedge Ratio Wet starchy produce plus a polished face Not a sharpening fault. Thin the blade, or accept it and use a taller blade to sweep. Dimples help slightly and fix nothing Partly — geometry only High force needed on soft produce (tomato, ripe pear) Tomato scale test, three readings Over 500 g with no visible chipping or rolling Apex has worn round through normal use Board material accelerating wear — bamboo, composite, or worse Full progression from 1000 grit, or 3000 grit alone if the edge is only lightly worn. Change the board if it is glass or stone At home, 10 minutes Orange rust spots Wipe with a dry cloth and look at what transfers Powdery orange residue, slightly raised, comes off on the cloth Moisture left in contact with carbon or low-chromium steel Acidic residue left on the blade, or storage in a damp drawer Cork or a wine cork with a little scouring powder, rinse, dry fully, then a thin film of camellia or food-safe mineral oil At home; deep pitting is permanent Grey or blue-grey patina spreading Wipe firmly with a dry cloth Nothing transfers; the colour is in the surface Normal stable oxide layer forming on carbon steel Contact with acidic food, which accelerates it Nothing. This layer protects the steel underneath. Polish it off only if you dislike it, knowing it will return Not a fault Scratched, hazy bevel after stoning Look at scratch direction under a light Scratches running at a different angle from the last grit used Grit contamination — coarse slurry carried onto a fine stone Skipping too far up the progression, so coarse scratches were never removed Rinse hands, blade and stone between every grit. Go back one grit and remove the coarse scratches fully before advancing At home, cosmetic Loose or rattling handle Grip the blade and twist the handle; look at the rivets and the scale join Any perceptible movement, or a dark line opening at the scale join Water ingress at the tang over years, usually from dishwashing Impact damage, or adhesive failure on a moulded handle Riveted handles can sometimes be re-peened by a cutler. Moulded and injection-bonded handles cannot be repaired Cutler only; often terminal Find your symptom, run the test in column two, and only act once the reading in column three confirms it. Two rows here — wedging and food sticking — are not sharpening faults at all, and sharpening them harder makes both worse. The row people misread most often is the fifth one. A knife that splits a butternut squash instead of slicing it feels blunt, gets sharpened, feels no better, gets sharpened again, and each round of sharpening makes the geometry marginally worse. The tomato test is what breaks that loop: if the apex reads under 250 g, the edge is not your problem.
The second most misread is the first row. Sharpening again does fix it — but only if the deburring step is done properly this time. Repeating the same sequence that created a wire edge produces another wire edge, and people conclude their steel is bad.
Rolled or chipped? Thirty seconds tells you
Both look like damage along the apex and both make a knife stop cutting. They have opposite causes and opposite fixes, so getting this wrong wastes an evening.
A rolled edge is steel that is still there, bent to one side. Under a light it shows as a continuous bright line that changes brightness as you tilt the blade, because you are catching a reflection off the folded-over face. A fingernail run across the edge catches noticeably on one side and feels smooth on the other. Rolling means the edge was too thin or too soft for the load it took.
A chip is steel that has gone. Under a light it shows as discrete bright spots rather than a line, and the shoulders of each nick look square and torn rather than folded. A fingernail catches equally from both sides. Chipping means the edge was too hard or too acute for the load it took.
So the corrective actions point in opposite directions. Rolling says go steeper by 2° per side, or use a softer touch on the board. Chipping says also go steeper, but the underlying instruction is different: keep this blade away from bone and frozen edges, because at 61 HRC or above it will never tolerate them regardless of angle.
One useful shortcut. If the damage recurs in the same place on the blade every time, it is not a steel problem — it is a habit. Most people who chip repeatedly are hitting the same point on the same board against the same hard thing.
When the edge is perfect and the blade is fat
This is the failure nobody warns buyers about and it accounts for a large share of expensive knives that get quietly replaced.
Sharpening removes steel from the bevel. That moves the apex up into the blade, and the blade gets thicker as you go up. Do it fifty times over eight years and the apex now sits where the steel is 0.7 mm thick instead of 0.3 mm. The apex is still razor keen. The wedge behind it has more than doubled, and every dense vegetable now splits ahead of the cut instead of parting around it.
Measure the spine thickness 5 mm above the edge with calipers and the situation is unambiguous. Under 0.35 mm is thin behind the edge and will fall through almost anything. Between 0.35 and 0.6 mm is a normal working German chef knife. Over 0.7 mm and you have a wedge with a sharp front on it.
The fix is thinning, and it is a different job from sharpening. You lay the whole primary bevel down on a coarse stone at roughly half the edge angle and work the flat until the thickness above the apex comes down, then re-establish the edge normally. It takes around forty minutes on a 400 grit stone, it removes a visible amount of metal, and it will change the look of the blade near the edge. It also transforms how the knife cuts in a way that no number of finer grits ever will.
Factory knives sometimes arrive in this state from new, particularly heavy forged Western chef knives that were ground for durability rather than for cutting. Thinning a new knife feels like vandalism and is frequently the single best thing you can do to it.
Matching angle to hardness
Measured HRC Safe minimum bevel °/side Typical factory bevel What happens below the minimum Microbevel needed? 54-56 (soft stainless) 18° 20-22° Edge rolls within a single board session No — thin the primary bevel instead 56-58 16° 15-17° Rolls under lateral load, e.g. levering through a squash Optional, 20° at one or two very light strokes 58-60 14° 15-16° Rolls on bone contact and cold, firm produce Helpful on a working knife 60-62 12° 14-16° Micro-chips rather than rolling Yes — 15-16° at two or three passes 62-64 (powder steel) 10° 12-15° Chips visibly on the first hard contact Yes — 15° minimum, no exceptions Zirconia ceramic Factory grind only ~20° inclusive Fractures. There is no home recovery Not applicable Check your knife's hardness band, then confirm your sharpening angle sits at or above the safe minimum. If you are already above it and the edge still fails, the fault is deburring or board material, not angle. A microbevel is a tiny secondary facet at a slightly steeper angle, put on with two or three passes at the very end of a sharpening session. It costs almost nothing in cutting performance and adds a surprising amount of edge stability on hard steel. On soft steel it is a waste of time — soft steel does not chip, it rolls, and a microbevel does not stop rolling.
There is a common misreading of this table worth heading off. A lower minimum angle for harder steel does not mean harder steel is more durable. It means harder steel can support a thinner geometry before it deforms. Its failure mode is worse when it comes, because chips need removed metal to fix and rolls do not.
Repairs to hand to somebody else
- Chips deeper than about 1 mm. Removing that much steel by hand on a 400 grit stone takes hours and almost always produces an uneven profile. A cutler on a belt grinder does it in minutes.
- A snapped tip. Reprofiling a broken tip means reshaping the whole front third of the blade so the new profile flows, and doing it by hand produces a knife with a strange, blunt-looking nose.
- Anything you have run through a pull-through carbide sharpener repeatedly. Those devices tear a scalloped, over-thinned edge that needs the whole bevel taken back before a stone can do anything useful.
- Bolster interference at the heel. Once the edge has retreated behind a full bolster you need the bolster ground back, and that is a grinder job with a real risk of ruining the blade's lines.
- Overheated steel — a blue or straw-coloured band near the edge after someone used a dry powered grinder. The temper is gone in that zone and the steel will never hold an edge again. Terminal, and there is no fix.
- Loose moulded or injection-bonded handles. Riveted scales can sometimes be re-peened. Bonded handles cannot be reopened without destroying them.
The habits behind most of these faults
Scraping prep off the board with the edge instead of the spine. This is the single most common cause of a rolled edge in a home kitchen, and it happens dozens of times a day. Flip the knife over. It costs nothing.
Cutting on glass, stone or a ceramic plate. Those surfaces are harder than the steel, so contact damages the apex rather than the board. One meal's prep on glass can undo a full sharpening session.
Leaving the knife in a bowl of water under a pan. Water at the tang join is what eventually loosens handles, and blades knocking together in a bowl chip tips.
Levering with the tip — prising apart a joint, opening a stubborn packet, popping a lid. Hardened steel has almost no strength in the sideways direction, which is precisely the direction levering loads it.
Buying a 62 HRC blade and treating it like a 56 HRC one. Hard Japanese steel is not more robust than German stainless; it holds an edge longer and tolerates less abuse. Choosing it means changing habits, not just changing knives.
Reaching for the honing rod when the tomato test reads 800 g. Past a certain point the apex is gone rather than bent, and no amount of realigning brings back steel that is not there. That is the moment for a stone.
Fix the habits before you buy anything. Most people who conclude their knife will not hold an edge have a perfectly capable knife, a burr they never removed, and a bamboo board.