For a 2x72 grinder, run ceramic in 36–120 grit for profiling, bevels and post-heat-treat cleanup, then move to aluminum oxide or flexible ceramic from 220 to 800 for finish work, and finish with a non-woven surface conditioning belt for satin. Zirconia is the value choice for handle material and light pre-heat-treat shaping. A 1x30 is a sharpening and small-work machine, not a bevel grinder — use it for edges, guards, and detail.
The belts you choose matter more than almost anything else on the grinder. A ceramic belt that cuts is worth three that don't, and the difference shows up in your finish, your heat control, and how many blades you get out of a shift.
Why 2x72 became the standard
Two inches is wide enough to grind a full bevel in one pass on most blade widths, and narrow enough to control on a contact wheel. Seventy-two inches gives you enough belt surface to absorb heat between passes — a shorter belt on the same wheels runs hotter and dies sooner because each square inch of abrasive comes back around more often.
The practical reason it stuck, though, is the ecosystem. Bader, KMG, Burr King, Reeder, AmeriBrade, TW-90 and the Grizzly G1015 all run the same belt. The 2x72 format is widely supported across abrasive types. Switch grinders and your abrasive inventory still works. Nothing else in knife making is that portable.
The real grit progression
This is the sequence most working shops actually run, stage by stage. Skipping more than one step costs you more time chasing scratches than it saves.
| Stage | Grit | Mineral | Why |
|---|---|---|---|
| Forge scale removal | 24–36 | Ceramic, or a used/expendable belt | Scale is harder than the steel under it and it kills fresh belts. Many smiths deliberately use a worn belt here. |
| Profiling the blank | 36–60 | Ceramic | Highest material removal per dollar. Zirconia works but you'll burn through more of them. |
| Rough bevel grinding, pre-heat-treat | 36–60 | Ceramic | Annealed steel cuts easily; the limit is your control, not the belt. |
| Refining bevels, pre-heat-treat | 80–120 | Ceramic or zirconia | Zirconia is fine here because the steel is soft. |
| Post-heat-treat cleanup | 60–120 | Ceramic only | Hardened steel work-hardens aluminum oxide and glazes zirconia fast. Ceramic is the only mineral that stays economical. |
| Finish grinding | 180–400 | Ceramic (flexible backing) or aluminum oxide | This is where the scratch pattern you'll live with gets set. |
| Hand-finish prep | 400–800 | Aluminum oxide, J-flex | Conforms to the blade, evens the scratch pattern, prepares for hand sanding. |
| Satin / final finish | Coarse to very fine | Non-woven surface conditioning | Blends, does not cut deep. Will not change your geometry. |
| Handle shaping (wood, micarta, G10) | 36–120 | Zirconia or aluminum oxide | Ceramic is wasted on handle material, which cuts easily and loads belts. |
| Handle contouring and finishing | 120–400 | Aluminum oxide J-flex | Slack belt work on a soft backing. |
Ceramic vs zirconia vs aluminum oxide, by stage
Premium ceramic grain is engineered to fracture under load, exposing new cutting edges as it wears. That is why it outlasts everything else on hard steel — the belt is constantly resharpening itself. It is the correct choice from first profiling through post-heat-treat cleanup, and it is the only mineral that makes economic sense on hardened blades.
Zirconia is tough and cuts well on soft steel, mild steel and non-ferrous. It is the right answer for pre-heat-treat shaping when you want to stretch your budget, for stock removal on annealed blanks, and for handle material. It does not have ceramic's self-sharpening behavior on hardened steel — it glazes instead.
Aluminum oxide is the workhorse for everything fine. It comes in the widest grit range and gives the most uniform, predictable scratch pattern, which is exactly what you want from 220 up. It is also the cheapest belt in the shop, and above 320 grit there is very little reason to pay for anything else.
Belt speed, and why ceramic needs pressure
Belt speed is measured in surface feet per minute. The formula is straightforward:
SFPM = RPM x 0.262 x drive wheel diameter (inches)
An 1800 RPM motor on a 4" drive wheel gives roughly 1,890 SFPM. A general-purpose starting point for steel is around 3,000 SFPM, and ceramic belts are commonly run between about 3,600 and 4,500 SFPM, with the higher end for coarse grits and the lower end for 80–120. Drop the speed substantially for wood, micarta and G10 — those burn.
Here is the part that costs beginners money: ceramic belts need pressure to work. The grain only fractures and renews itself when it is loaded. Feather a ceramic belt the way you'd feather an aluminum oxide belt and the grain rounds over instead of breaking, the belt glazes, and you conclude ceramic is overrated. Lean on it. If you are not making chips and the belt is not doing visible work, you are not pushing hard enough.
The corollary: back off on aluminum oxide and on anything above 320. Fine grits cut by scratching, not fracturing, and excess pressure just generates heat and rolls your edges.
When J-flex earns its place
A J-weight flexible backing is thinner and more pliable than the X-weight cloth used on grinding belts. That flexibility is a liability on a platen — it won't hold a flat bevel — and an asset everywhere else.
Plunge lines. A J-flex belt run on a slack section or a soft contact wheel will follow the transition instead of cutting into it. It cleans up and blends a plunge without deepening it, which is the single most common place a nearly finished knife gets ruined.
Handle contouring. Slack belt work on a handle is where J-flex is not optional. The belt wraps the palm swell and the finger grooves and leaves an even surface. An X-weight belt bridges the contours and cuts flats into them.
Ricasso, spine and choil. Anywhere you need the abrasive to conform to a shape rather than impose one.
We run AO J-flex in the fine grits for exactly this work. It is a different tool from a grinding belt, not a cheaper version of one.
Surface conditioning belts for finishing
Non-woven surface conditioning belts are an open nylon web with abrasive bonded through it. They do not remove meaningful stock — they blend. That is the point.
Use them after your last coated grit to produce a consistent satin finish, to knock down a scratch pattern that is slightly uneven, or to finish stainless where you want a uniform directional grain rather than a mirror. Because the web is open, they run cool and do not load on soft metals the way a coated belt does. They will also not correct geometry, so do not reach for one to fix a wavy bevel.
Shop surface-conditioning abrasives
1x30 for sharpening and small work
A 1x30 is not a small 2x72. It is a different machine with a different job.
Where it earns its place: putting and maintaining a working edge, with a leather or fine belt for the final pass; shaping and fitting guards and bolsters; detail work on small blades and kitchen knives; deburring; sharpening chisels, plane irons and other shop tools.
Where it does not: grinding bevels on anything substantial. One inch of belt is not enough to establish a flat bevel on a blade of any width, the small wheels heat the belt quickly, and there is not enough belt surface to carry the heat away. People do it, and the results show.
For 1x30 sharpening, a short progression works better than a long one: a coarse belt to set the bevel, 220–400 to refine, then a fine or non-woven belt to finish. Keep the speed and pressure low on a hardened edge — a 1x30 will draw the temper out of a thin edge faster than you expect.
Belt life and cost per knife
Cost per belt is the wrong number. Cost per knife is the right one.
A cheap aluminum oxide belt on hardened steel might cut well for two minutes, glaze, and then spend the next ten minutes generating heat instead of removing metal. You will use four of them to do what one ceramic belt does, and you will have spent forty minutes of your labor plus the risk of drawing temper. The belt was cheaper. The knife was more expensive.
Practical ways to actually reduce cost per knife:
- Don't skip grits. Jumping 60 to 220 means the 220 belt does the 120 belt's work, badly and slowly, and wears out doing it.
- Reverse butt-spliced belts. A bi-directional belt that's loading up can be pulled off and reinstalled the other way around. You get meaningful cut back for free.
- Keep a scale belt. Dedicate a worn belt to forge scale so it never touches a fresh one.
- Match mineral to stage. Ceramic on hardened steel, zirconia on soft steel and handles, AO above 220. Using ceramic for handle sanding is money on the floor.
- Run the right speed. Belts die from heat as often as from wear.
Track how many blades you finish per belt at each stage for a month. That number will tell you more about what to buy than any spec sheet.
Sizes and materials we make
We convert and splice sanding belts through our Anaheim operation. 2x72 and 1x30 are stock sizes; so are 1x42, 2x42, 2x48 and a lot of others — dozens of widths and lengths across the catalog. Minerals: premium ceramic, zirconia, aluminum oxide in cloth and paper, silicon carbide, AO J-flex, and non-woven surface conditioning, from 24 grit to 800.
If you built your own grinder and need a length nobody stocks, we make custom belts from 1/4" to 60" wide and up to 400" long with low minimums; ask for the current lead time and a contact us for the current quote time.
Request a custom sanding belt quote
Frequently asked questions
What grit sanding belt should I start with for knife making? 36 or 60 grit ceramic for profiling and rough bevel grinding. If you're removing forge scale first, use 24–36 on a belt you've already worn out — scale destroys fresh abrasive.
What is the best belt for knife making, ceramic or zirconia? Ceramic for anything on hardened steel and for heavy stock removal, because the grain fractures and resharpens itself under load. Zirconia for soft or annealed steel and for handle material, where it costs less and works nearly as well.
How fast should a 2x72 belt grinder run? Around 3,000 SFPM is a safe general setting for steel. Ceramic belts generally perform best in the 3,600–4,500 SFPM range, faster for coarse grits and slower for 80–120. Slow down substantially for wood, micarta and G10, which burn.
Why isn't my ceramic belt cutting? You're probably not pushing hard enough. Ceramic grain only renews itself when it's loaded; light pressure rounds the grain over and glazes the belt. Increase pressure until the belt is visibly removing metal.
What is a J-flex belt for on a knife grinder? J-weight flexible belts conform to shapes instead of imposing flats. Use them on slack belt or a soft wheel for blending plunge lines, contouring handles, and finishing the spine, ricasso and choil. Don't use them on a flat platen — they won't hold a flat bevel.
Can I grind bevels on a 1x30? Not well. A 1x30 doesn't have the belt width to establish a flat bevel or the belt surface to shed heat. Use it for sharpening, guards and bolsters, small blades and detail work.
How long should a 2x72 belt last? It depends entirely on mineral, stage and how you run it, so measure it in blades per belt in your own shop rather than in minutes. Ceramic on hardened steel will outlast aluminum oxide by a wide margin; track your own numbers per stage for a month and buy accordingly.
What grits do I need for a full knife from start to finish? A typical working set: 36 and 60 ceramic for profiling and rough bevels, 80 and 120 ceramic for refining and post-heat-treat, 220 and 400 for finish grinding, 600–800 aluminum oxide or J-flex for hand-finish prep, and a non-woven surface conditioning belt for satin.