Knife Bevel Grinding: How to Get Clean, Even Bevels and Crisp Plunge Lines

01 What is a bevel, actually?

A knife has two bevels, and beginners constantly conflate them.

The primary bevel is the big one — the ground face that takes the blade from full stock thickness at the spine down to nearly nothing at the edge. This is what you make on the grinder. It's what people mean by "a flat grind" or "a hollow grind."

The edge bevel (or secondary bevel, or micro-bevel) is the tiny one at the very edge, put there when you sharpen. It's typically a fraction of a millimetre wide.

Two different angles, two different operations, two different stages of the build. Everything below is about the primary bevel.

02 Why is my bevel angle so much shallower than I expected?

Because it is, and this surprises almost everyone.

You sharpen at 15 degrees per side. So people assume they're grinding a bevel somewhere in that neighbourhood. They're not — a primary bevel on a typical knife is around two degrees per side.

The geometry is simple. You're removing material from the full stock thickness down to your target edge thickness, spread across the height of the bevel. Per side you remove half of that difference. So sin(θ) = ((stock thickness − edge thickness) / 2) / bevel height.

Bevel height is measured from the edge up to where the grind stops. Strictly, sin wants the width of the finished ground face and tan wants the perpendicular height you actually scribe on the flat blank. At these angles the difference is under 0.03° anywhere in the table below — use whichever you can measure.

Run real numbers on a standard first knife — 1/8" stock, 0.030" left at the edge before heat treat, a 1.25" tall blade ground full height — and you get 2.18 degrees per side. A 4.4 degree included angle.

Why this matters practically: at two degrees, a one-degree error in how you're holding the blade is a fifty percent error in the grind. That's why bevels wander, why one side ends up taller than the other, and why a jig helps so much. You are not making a coarse adjustment — you're holding a very shallow angle very consistently for several minutes.

A note on the "two degrees" figure: it's specific to a full-height grind on 1/8" stock. The table below runs from about 1° to 6°. The general claim is that primary bevels are single digits, not the fifteen degrees you sharpen at — both are measured from the blade's centreline, so they're directly comparable.

And on thin stock: people say thin stock is less forgiving, and it is, but not for the reason usually given. A given angular error removes the same amount of extra material regardless of thickness. What changes is the budget — on thin stock there's simply less material between a wobble and a hole, and it warps more in the quench.

03 What grind should I choose?

The grind is the cross-sectional shape of the primary bevel. Five show up in practice.

For your first several knives: flat grind. It's the most forgiving to learn, the easiest to hand sand (a flat backer chases a flat surface), the easiest to sharpen for the rest of the knife's life, and the easiest to diagnose when it goes wrong — a flat surface visibly shows you where it isn't flat.

Hollow grinds look fantastic and cut very well, but there's a constraint most guides never mention.

04 Will my contact wheel cut through the blade?

This is the hollow grind trap, and it catches people.

When you grind against a contact wheel, the wheel's curvature is cut into the blade. The depth of that hollow is fixed by the wheel diameter and the height of the grind — and it has nothing to do with how carefully you're working. If the geometry says the wheel removes more material than you have, it removes more than you have, and you grind a window through your blade.

The thing almost nobody says out loud: the thinnest point is not at the edge. The arc has to run from the edge — already down at your 0.030" — up to the grind line at full stock thickness. That arc bottoms out somewhere in the middle, typically a third to a half of the way up the bevel. So the blade goes translucent, and then holes through, in the middle of the bevel, while the edge is still exactly where you left it. People discover this by holding the blade up to a light and seeing daylight through it.

Once you fix the wheel radius and the two endpoints, the arc is completely determined — there's no technique that changes it.

Read the near-misses as failures too. A 1/8" blade on a 12" wheel at 1.25" leaves 0.0035" — about the thickness of a sheet of paper. That is not a knife, it's a warning.

05 How tall can I hollow grind?

Rather than working out whether you'll survive, work out how tall you can go. The useful limit is the height at which the thinnest point drops below your edge thickness — past that you're thinner mid-bevel than at the edge, which is structurally silly even when there's still steel there.

h_max = √( R² − ( R − (t − e)/2 )² ), where R is wheel radius, t is stock thickness, and e is edge thickness.

These heights are smaller than most people expect, and that's the point. A hollow grind on a common 8" or 10" wheel is a saber-height grind, not a full-height one. The tall, elegant hollow grinds you see on production knives come off 14-inch-and-up wheels or dedicated hollow grinding machines.

The rule that falls out of it: small wheels want short grinds; tall grinds want big wheels. If you own an 8" wheel and want a tall grind, grind flat instead.

06 How high should the grind go?

Grind height is usually described as a fraction of blade height.

For a first knife, grind high flat. Leave a band of maybe 1–2mm of unground flat below the spine. Feathering a grind out to a true zero at the spine is genuinely difficult and it's the thing that most often produces a wavy grind line at the top of the blade.

• ◆ Full flat / full height — spine to edge, no flat left. The best slicer. Standard on kitchen knives.
• ◆ High flat — grind stops just below the spine, leaving a thin flat band. Very common on general-purpose knives; slightly stronger spine, much easier to grind than a true full flat because you're not trying to feather out at the very top.
• ◆ Saber, roughly 50–70% — a distinct flat band on the upper blade. Stronger, thicker behind the edge, common on hard-use and outdoor knives.
• ◆ Scandi, roughly 20–30% — low and steep, sharpened at the primary bevel with no secondary.

07 What do I do before I touch the belt?

Half of a good bevel is set up before the machine turns on. By the time you start the machine you should have: two scribe lines at the edge, a grind height line on each face, a plunge mark, and a blade you've measured. Now the grinding is just hitting targets you can see.

Get the bit size right for the centerline: the band between the lines equals stock thickness minus bit diameter, so use a bit roughly your edge thickness smaller than your stock. A 3/32" bit on 1/8" stock leaves a 0.031" band — near enough perfect. A bit the same diameter as your stock scribes dead centre and gives you a zero band, which is exactly what you don't want. Treat the lines as a visual guide and still confirm with calipers.

• ◆ Flatten the blank. Bar stock isn't flat and isn't parallel. Grind or surface a few passes off each face, then check with a straightedge and calipers at four points. Every bit of thickness variation you leave here shows up later as a bevel that won't come even — and you'll blame your hands for it.
• ◆ Deburr and clean up the profile. File the outline fair before you grind bevels, not after. A lumpy profile makes a lumpy grind line.
• ◆ Blue the whole blade with layout dye. You need to read your progress, and marker rubs off.
• ◆ Scribe the edge centerline. Lay the blade flat on the bench. Hold a drill bit flat on the bench beside it and drag it along the edge — the widest point of the bit sits at half its diameter, so it scribes a line onto the narrow edge face at that height. Flip the blade and repeat, and you have two lines with a band of steel between them.
• ◆ Scribe the grind height line on both faces with a height gauge or a square and a scribe. Two lines, same height, both sides. This is your target for where the grind stops.
• ◆ Mark the plunge with a square.
• ◆ Check your edge thickness target. Leave 0.025"–0.035" before heat treat. Grinding to zero pre-heat-treat gets you a warped, decarburised, possibly cracked blade.

08 Should I use a jig or grind freehand?

The honest answer: use a jig for your first ten knives, then learn freehand if you want to.

Freehand grinding is a genuine physical skill that takes dozens of blades to develop. There's a romance around it that produces a lot of bad advice, mostly from people who forgot how long it took them.

Here's the practical case for the jig that most people miss: a jig teaches you what a correct bevel looks and feels like. If you grind freehand from day one, you're trying to learn the motion and recognise the target simultaneously, with no reference for either. Grind twenty good bevels with a jig and your eye is calibrated. Then freehand has something to aim at.

The options, roughly cheapest first: a shop-made file-guide-and-block setup; an angle-adjustable work rest (the most versatile, and useful for a hundred other things); a bubble level jig like the Bubble Jig, which clamps to the tang and lets you hold a set angle by watching a level; and full tilting-table jigs.

If you buy one thing, buy an angle-adjustable work rest. It's not blade-specific, it works for sharpening and squaring and handle work, and it's the foundation for most other methods.

09 How do I set the angle on a jig?

Work backwards from the geometry. Then stop grinding 1–2mm short of your scribed line. You want material left for finishing grits and for the post-heat-treat cleanup. Grinding to the line at 36 grit means the line is gone by the time you're at 400.

• ◆ Measure your stock thickness with calipers.
• ◆ Decide your edge thickness — 0.030" is the sensible default.
• ◆ Measure your bevel height — from the edge up to your scribed grind line, measured along the face.
• ◆ Compute the angle: θ = arcsin(((thickness − edge) / 2) / height), or use one of the free online bevel calculators.
• ◆ Set the rest to that angle. The precise way: zero your digital inclinometer on the platen first — otherwise it's reading against gravity and your target number means nothing. Then clamp the blade spine to a square block (a 123 block works), sit the inclinometer on the blade face, and adjust the rest until it reads your target. Approach the setting from below — start the rest low and come up — so you're always taking up backlash in one direction.
• ◆ No inclinometer? Add 90° to your target and set that angle between the rest and the platen with a protractor.

10 How do I actually grind freehand?

When you're ready, the mechanics matter more than the mystique.

Your hands don't set the angle — your body does. Set your stance, lock your wrists and elbows, and move the blade by shifting your whole upper body. If you're steering with your fingers, the angle changes constantly. This is the single biggest thing beginners get wrong.

Grip: blade held with the spine against your fingers, edge trailing, both hands on the blade — one guiding, one applying pressure through the area currently in contact. Pressure comes from the hand behind the contact point, not from the hand holding the tang.

The pass: start at the plunge, move toward the tip in one smooth continuous motion, and lift off. Don't stop mid-blade. Don't reverse direction inside a pass. As you approach the tip, two things change at once — the blade gets thinner and the belt is removing material faster — so lighten pressure progressively.

The belt: grind on the flat platen with the belt tracking consistently. Use a fresh belt. A worn belt cuts unevenly across its width, which puts a ridge in your bevel that you'll spend an hour chasing.

Practice on scrap first. Cut six blanks from cheap 1084 and grind bevels on all of them with no intention of finishing any. It's the fastest way to learn and it costs about thirty dollars.

11 How do I get crisp plunge lines?

The plunge is where the bevel begins and the ricasso ends. Matched, crisp plunges are the first thing an experienced eye checks — they're a much better tell of skill than the bevel itself.

Whichever approach you use, the rule is: match them. Sight along the spine with the blade edge-on and look at both plunges together. They should start at the same distance from the ricasso and have the same shape. A blade with mismatched plunges reads as sloppy no matter how good the bevels are.

Cheat: scribe the plunge on both sides from the same reference — measure from the tip or from a pin hole, not from the butt of the tang, which you may still reshape.

• ◆ Hard stop with a file guide. Clamp a file guide across the blank at the plunge line and let it physically stop the belt. You get a dead-square plunge, repeatable, on the first attempt. This is the beginner answer and it produces excellent results. Downside: the plunge comes out perfectly square, which some people find visually stark. Get the right guide — a plain hardened guide is fine for filing, but a ceramic belt will happily cut hardened tool steel. If you're stopping a belt against it, buy one with carbide inserts (Jantz, Bill Behnke and others sell them for exactly this). Expect it to eat belts either way — that's the trade.
• ◆ Push past and blend. Run the grind past where you want the plunge, then come back and blend the transition by hand or with the corner of the belt. Produces a softer, sweeping plunge. Harder to keep symmetric — this is where most mismatched plunges come from.
• ◆ Radiused plunge on the wheel. Grind the plunge with the small wheel or the rounded edge of the platen, giving a deliberate radius. Looks intentional and modern, and it's more forgiving than a square plunge because a radius hides small mismatches that a hard line would show.

12 How do I keep both bevels symmetric?

Symmetry is a process problem, not a talent problem. Four habits solve almost all of it.

• ◆ Count your passes. One pass on this side, one pass on that side, and keep a running count out loud or in your head. Not "about the same" — the same number. This is the single habit that most separates even bevels from uneven ones.
• ◆ Alternate sides at every grit change, and start the new grit on the side you finished the last one on. Otherwise a systematic bias builds up through the progression.
• ◆ Measure, don't look. Put calipers on the edge thickness at four points along the length — near the plunge, two in the middle, one near the tip — for both sides. Write the numbers down. Your eyes will tell you the bevels match long before they do, because the brain is extremely good at seeing symmetry that isn't there.
• ◆ Re-dye between grits. A fresh coat of layout dye makes it obvious which parts of the bevel you're actually touching. You will be surprised.

13 Why does my grind look wrong?

Diagnose the cause before you grind again. Every extra pass on a bad bevel makes it worse. Grinding only removes metal, so a wavy grind line can only move up — bring the low sections up to the height of the wave's highest point. That raises your grind line overall, which is exactly why you stop short of the scribe.

14 How do I finish the bevels?

The finishing progression is where the bevel stops being a shape and starts being a surface.

Before heat treat: work down to roughly a 400-grit equivalent — in belts, that's stepping 36 or 40 ceramic → 60 → 120 → Trizact A65 → A45. Each grit's only job is removing the scratches from the previous one, so change your scratch direction slightly at each step and keep going until no coarser scratches remain. Stop while the edge is still 0.030".

After heat treat: clean the scale off with a fresh 120 belt, very light pressure, and a water dunk between every single pass — this is where hardened blades get ruined. Then run the belts back up the progression — 120 → 220 → A65 → A45 — before you pick up sandpaper. Skipping straight from a 120 belt to hand sanding means chasing 120-grit scratches by hand, which is punishing and nobody actually does it.

Bring the bevels close to a zero edge as you go, leaving a hair of flat — around 0.005"–0.010", roughly two sheets of printer paper — so the edge doesn't chip or curl on the belt.

Then hand sand. A hard flat backer, one direction per grit, 220 → 320 → 400 → 600. This is where the flat grind pays you back: a flat backer on a flat surface tells you honestly where the low spots are.

15 How should I actually practise this?

Deliberately, on material you don't care about. Six practice blanks cost roughly the price of a takeaway meal and will save you a heat treat fee and a weekend. Every experienced maker has a box of these somewhere.

• ◆ Buy a 24" bar of 1084 and cut six identical blanks. Same size, same simple shape.
• ◆ Grind all six with a jig. Don't finish them. Measure each one and write down where it went wrong.
• ◆ Grind three more freehand, comparing against the jig-ground reference you now have in hand.
• ◆ Only then grind the blank you intend to make into a knife.

16 The short version

A primary bevel is single digits — about two degrees on a full-height grind on 1/8" stock — not the fifteen you sharpen at. Precision matters more than force. Flatten the blank, scribe the edge centerline and the grind height on both faces, and leave 0.030" at the edge. Grind flat on the platen with a jig, high flat rather than full flat, count your passes, alternate every pass, and measure with calipers rather than trusting your eye. Use a file guide for the plunges and match them by sight along the spine. Stop 1–2mm short of your scribed line at coarse grit. If it goes wrong, diagnose the cause before you grind again — every extra pass on a bad bevel makes it worse.

And grind six practice blanks first.

Frequently Asked Questions

Questions from the Counter

Do I grind bevels before or after heat treat?

Rough them in before, at 36 through A45 grit, leaving 0.030" at the edge. Finish them after. Grinding a hardened blade is slower and generates far more heat, so you want most of the material gone while the steel is soft.

Can I grind bevels with just a file?

Yes — with a file guide clamped at the plunge and a coarse file, it's a real method and it's how knives were made for centuries. It's slow and hard work, but it's extremely controlled, and it's a legitimate way to make a good knife with no grinder at all.

What edge thickness should I leave before heat treat?

0.025"–0.035" for carbon steel. Thinner risks warping and cracking in the quench; much thicker means a lot of slow, hot grinding on hardened steel afterwards.

Why does my bevel look fine until I put dye on it?

Because dye is honest and steel is shiny. Reflective surfaces hide low spots extremely well. Re-dye between grits and trust what it shows you.

How long should bevel grinding take?

On a first knife, an hour or two per side is normal, and slow is correct. The goal at this stage is control, not speed. Speed comes from confidence, and confidence comes from having done it twenty times.

Should the grind be perfectly symmetric?

The two bevels should be, yes. But note that some knives are deliberately asymmetric — many Japanese kitchen knives are ground with different angles on each side for specific cutting behaviour. That's an advanced choice, not an excuse for an accident.

From the Page to the Palm

Reading about steel is one thing. Holding it is the other.

Every claim in this guide is backed by a piece on the shelf — knife-making tools, belts, and steel, named and priced plainly in CAD.

More in This Guide