A soft bond for hard concrete? That sounds backwards.

It is backwards from the guess, and the reason is that the bond is not doing the cutting. The diamonds do the cutting. The bond is the metal matrix holding them, and its job is to wear away at the right speed so that blunt diamonds are released and fresh ones are exposed. Hard, dense concrete barely abrades that matrix. A hard bond on it therefore never wears, never releases a fresh diamond, and the segment polishes over and stops cutting. A soft bond erodes on its own, which is exactly what a slab that will not erode it for you requires. Soft, abrasive concrete does the opposite: it grinds the matrix away far faster than the diamonds are used up, so a soft bond is thrown away half worn and a hard bond is what survives long enough to earn its cost.

How do I tell whether a slab is hard or soft before I start?

Not from the mix design, and this is the part that catches people out. Compressive strength and surface hardness are different properties. A high strength slab that was power troweled hard has a dense burnished skin, and a lower strength slab that was finished wet can be soft enough to scratch with a screw. Three field checks get you close. Score the surface with a masonry screw or a nail: if it marks easily and throws powder, treat it as soft. Look at what the floor has been through, because a densifier, a cure and seal or years of traffic all harden the top. Then settle it the only way that is actually reliable, with a trial patch: run a small area and look at the segments and the swarf after a few minutes.

What is the difference between bond and grit?

They are two independent dials and they get conflated constantly. Grit is the size of the diamond particles, and it sets how coarse a scratch the tooling leaves, which is what puts you at one profile rather than another. Bond is the hardness of the metal matrix, and it sets whether the tooling keeps cutting at all. The practical consequence is that a wrong grit gives you the wrong finish and a wrong bond gives you no progress. You can be running exactly the right grit for the profile you were asked for and still be getting nowhere, because the segments have glazed and no diamond is presented to the floor.

My segments went shiny and stopped cutting. What happened?

That is glazing, and it is the classic hard bond on hard concrete failure. A healthy segment looks matte and slightly rough, with diamond visible in the face. A glazed one looks polished and feels smooth, the machine starts to ride rather than bite, the cut slows to nothing and the head runs hotter than it should. You can usually re-open the face by dressing it, which means running the machine over something deliberately abrasive to strip metal back off the segment and expose diamond again. Treat that as a diagnosis rather than a fix: if it glazes again within the next pass, the bond is too hard for that slab and the tooling has to change.

Do bond colors and numbers mean the same thing across brands?

No, and it is worth being blunt about it because it is a common way to buy the wrong set. There is no industry standard scale for bond hardness. Manufacturers use their own numbering, their own color coding and their own descriptions, and a color that means soft in one range can mean something else entirely in another. So a rule copied from a forum post or from a supplier you no longer buy from does not transfer. What does transfer is the underlying relationship in this guide and the evidence from your own segments. Ask the supplier which of their bonds is intended for the slab you are describing, and then confirm it on a trial patch before committing to a full set.

Will one bond do the whole job?

Often not, and the reason is that the slab is not one material all the way down. The troweled surface is denser than the concrete a few millimeters beneath it, and anything applied to it since has hardened it further. So the first pass can behave like hard concrete and the passes after it like softer concrete, on the same floor, with the same machine. Floors also vary across their own area. An aisle that has been driven on for twenty years is not the bay beside it, a section poured on a different day is not the rest of the slab, and a patched area is its own material. If the cut changes character partway across a floor, that is information rather than a malfunction, and it usually means the tooling that was right at the door is no longer right at the back wall.

Diamond Tooling: Bond vs Concrete Hardness

Hard concrete takes a soft bond. Soft concrete takes a hard bond. The rule is backwards from the guess because you are not matching the diamond to the slab, you are matching the metal holding it.

Quick answer

Hard concrete takes a soft bond. Soft concrete takes a hard bond. The rule reads backwards because the bond is not what cuts. The diamonds cut, and the bond is the metal holding them, which has to wear away at the right speed to keep releasing blunt diamonds and exposing fresh ones. A slab that will not wear the metal needs a bond that gives way on its own. A slab that eats the metal needs one that resists. You are not matching tooling to how strong the concrete is, you are matching it to what the concrete does to the metal, and those two are close to opposites.

What a diamond segment is actually doing

A metal bond segment is industrial diamond suspended in a sintered metal matrix. Only the diamonds sitting proud of that matrix are in contact with the floor, and they are the only part doing any work. They are harder than anything in the slab, so they do not really wear down. What happens instead is that they go blunt, they fracture, and eventually they are torn out.

That is not a fault. It is the design. As the matrix around a spent diamond erodes, the diamond falls away and the next layer down is presented to the floor. A segment that is cutting well is one that is being consumed at a steady, controlled rate, which is why diamond tooling is a consumable rather than a tool you own. The whole selection problem is that the erosion has to happen at the right speed, and the slab is the thing setting that speed.

The rule is backwards from the guess

Ask most people to match tooling to a hard floor and they reach for something hard. It is the wrong instinct here, and following it is the most common reason a grinder spends an afternoon achieving nothing.

Hard, dense concrete takes a soft bond

A hard slab barely abrades the matrix, so nothing wears the segment for you
A hard bond on it therefore never releases a spent diamond, and the face polishes over
A soft bond erodes readily on its own, which is the only way fresh diamond keeps reaching the floor
It wears faster in absolute terms, and that is the price of cutting at all

Soft, abrasive concrete takes a hard bond

A soft or sandy slab is effectively grinding the tooling as much as the tooling is grinding it
A soft bond is stripped away far faster than the diamonds in it are used up
You throw away segments that still have most of their diamond left, which is money on the floor
A hard bond resists that erosion long enough to get the diamond’s full life out of it

Said in one sentence: the bond has to supply whatever wear the slab does not. That is why the relationship inverts, and it is worth learning as a mechanism rather than as a rule, because a mechanism survives being half remembered on a job and a rule does not.

The two ways it goes wrong, and how each one looks

Both failures are visible within minutes if you look at the segments rather than at the floor, and they look nothing like each other. That makes this one of the few decisions in preparation you can correct on the day instead of discovering months later.

Glazing: the bond is too hard

The segment face goes shiny and smooth instead of matte, and you can no longer see diamond in it
The cut slows to almost nothing while the machine keeps running at the same speed and sound
The head starts to ride over the surface rather than bite into it, and it runs hotter than it should
There is very little swarf coming off for the time you are putting in

Premature wear: the bond is too soft

Segment height drops visibly over a single session, far faster than the job should have cost
The set is finished while the diamond in it is not, so you paid for diamond you never used
Cost per square foot climbs steeply and does not match anything you quoted
The cut can be aggressive enough to leave marks you then have to spend another pass removing

A glazed face can often be re-opened by dressing it, which means running the machine deliberately over something abrasive to strip metal back and expose diamond again. Treat that as a diagnosis rather than a repair. If it glazes again on the next pass, the bond is wrong for that slab and no amount of dressing changes what the tooling is.

Bond is not grit

These are two independent choices and conflating them wastes a lot of tooling. Grit is the size of the diamond particles. It decides how coarse a scratch the tooling leaves, which is what puts you at one surface profile rather than another. Bond is the hardness of the metal holding those particles. It decides whether the tooling presents any diamond to the floor at all.

So the two failures are different in kind. The wrong grit gives you the wrong finish, which you can see and correct with another pass. The wrong bond gives you no progress, which looks like a slow machine, a tired operator or a hard floor, and is none of those things. You can be running exactly the correct grit for the profile the data sheet specified and still be getting nowhere.

The other bond decision on this site is a different axis again. On a polished floor, metal bond and resin bond describe what the tooling is for, removal against refinement, rather than how fast the matrix erodes. Both questions exist at once on a polishing job: which family, and then how hard within it.

Reading the slab before you commit to a set

The single most useful correction here is that compressive strength is not surface hardness. They are different properties and they come apart regularly. A high strength slab finished hard has a dense, burnished skin. A lower strength slab finished wet can be soft enough to score with a screw. The number on the pour record tells you what the concrete can carry, not what it will do to your tooling.

Score it. Drag a masonry screw or a nail across the surface. An easy mark that throws powder says soft. A screw that skates and leaves a faint line says hard and burnished
Read the floor’s history. A densifier, a cure and seal, or twenty years of traffic all harden the top. So does a slab that was troweled late and hard
Expect the surface and the body to differ. The troweled skin is denser than the concrete a few millimeters under it, so the first pass and the passes after it can want different things
Expect the floor to differ from itself. A driven aisle, a bay poured on another day and a patched section are three materials, and the cut changing partway across is information rather than a fault
Settle it with a trial patch. Run a small area, then look at the segments and the swarf. Nothing else in this list is as reliable, and it takes minutes
Do not trust a color chart across brands. There is no industry standard bond scale. Numbers and colors are each manufacturer’s own, so a rule that worked with one supplier does not carry to the next

What else changes how aggressive the tooling is

Bond is the main dial but it is not the only one, and two sets described identically can behave differently on the same floor because of the rest of the setup.

Segment count and contact area. The machine’s weight is divided between however many segments are touching the floor. Fewer segments means more pressure on each one, a more aggressive cut and a shorter life. More segments spreads the same load, cuts more slowly, lasts longer and leaves a finer result. That is why changing to a fuller tool head can look like a bond change and is not one.

Machine weight and head pressure. The same bond under a heavier machine behaves as if it were softer, because there is more force driving the matrix into the floor. A bond that is correct on a small machine can glaze on a bigger one and the other way round, which is a real reason a set that worked for somebody else does not work for you.

Wet or dry. Water cools the segments and flushes the swarf away, which changes both wear rate and cut. Most coating preparation is run dry with extraction, because the slab has to stay dry and the dust has to be captured, and the prep pillar covers where that sits in the sequence.

The rate you plan with assumes the tooling is cutting

A glazed set does not slow the machine down, it slows the floor down. Plan the hours at a rate you have actually measured on this slab, and re-run it if the cut changes character partway through.

tradeepoxy.com/calculators/grinding-production-calculatorStatic preview

A trial patch gives you a rate for this floor rather than a rate for floors in general.

Open the live calculator →

What the wrong bond costs

It lands in two different places on a quote, which is part of why it stays invisible. The obvious one is tooling. A soft bond on an abrasive slab is thrown away with most of its diamond unused, and diamond is the largest consumable on a prep job. The grinding cost calculator prices that from your own set cost and your own lifespan, and a bond mistake shows up there as a lifespan that never matches what you assumed.

The less obvious one is hours. A glazed set does not announce itself: the machine sounds the same, moves at the same speed and produces almost nothing. Whatever rate you planned the job at is now wrong, and it is wrong for a reason that never appears in a schedule. Production rates by pass covers what else moves that number once the tooling is right.

Both of those are why the trial patch is worth the fifteen minutes. It is the only step in this decision that turns an assumption about a slab into a measurement of it, and it happens before you have committed to a full set of tooling or a price.

Where to go next

Tooling selection sits inside the profile gate, and how to prep concrete for epoxy covers all four gates and the order the work goes in. For what the grit half of the choice is aiming at, CSP explained covers the scale and why the number is a requirement rather than a preference.

For the other two ways to open a slab and where each one wins, grinding vs shot blasting turns on reach and grinding vs acid etching turns on whether you can verify what you produced. And for the polishing side of the bond question, the grit sequence covers where metal hands over to resin.