5 Measures to Prevent Cracks in Cemented Carbide Tools
Sharpening cracks in cemented carbide tools are rarely a mystery. They come from heat, vibration and edge stress that build up during grinding. This page explains where cracks in cemented carbide tools actually start, which parameters keep them away, and when the substrate is already past saving.

In this article
- 1
- 2
- 3
- 4
- 5
- 6
Why cemented carbide cracks instead of bending
Cemented carbide is tungsten carbide grains held together by a cobalt binder. It has almost no plastic range. A high-speed steel edge under overload will bend, then work-harden, and you can usually rescue it. Carbide reaches its elastic limit and fractures. That is the whole reason cracks in cemented carbide tools look sudden when the grinding history says nothing went wrong.
The binder is the weak link. Cobalt softens around 600–800 °C in the grinding zone, so the grains no longer have support exactly when thermal expansion is pushing them apart. Cooling from that state adds tensile stress at the surface. A 50 μm deep crack on a 0.8 mm micro end mill is not cosmetic. It is a crack that will run under load.
Substrate grade sets the baseline. Fine-grain grades with 6–10% cobalt resist edge chipping better than coarse grades at the same hardness. Submicron grades hold a sharper edge but tolerate less mechanical shock. Picking a grade harder than the job needs is one of the quiet causes of grinding cracks, because the wheel has to work harder to remove the same volume.
The practical consequence: you cannot inspect your way out of a bad grind. Crack prevention sits in the setup, not in the final visual check.
Heat is the first cause, and it is measurable
Most cracks in cemented carbide tools begin as grinding burn. The wheel rubs instead of cutting, the cobalt binder heats, and the surface layer goes into tension as it cools. You can see the evidence without a microscope: a straw or blue tint on the flank, a shiny smeared patch, or sparks that turn from orange to bright white.
Grinding ratio tells you where you are. A dull diamond wheel removes less material per unit of heat, so the operator pushes harder, and the cycle repeats. Dressing the wheel on a fixed schedule costs minutes. A scrapped batch of 200 inserts costs the whole shift.
Depth of cut per pass matters more than total stock. Keep each pass light, 0.005–0.02 mm for finish passes on carbide, and let the wheel cut rather than rub. Flood coolant aimed at the contact point, not at the fixture.
If the part is warm to the touch when it comes off the wheel, the surface already went through a thermal cycle. Warm is not the same as cracked, but it is the same road.
Vibration and clamping decide the second half of the risk
A carbide edge is brittle in tension, and vibration loads it in tension thousands of times per minute. The source is usually the setup, not the wheel. Long overhangs, thin webs, and a tool held in a three-jaw chuck instead of a collet all let the edge oscillate against the abrasive.
Negative rake hanging, or supporting the blade on both faces instead of one, spreads the contact area and drops the impact load. It is a simple change. It is also the difference between a 20-piece run and a 200-piece run on thin blades.
Wheel balance and spindle runout belong on the same checklist. A wheel with 0.02 mm of runout will strike one flute harder than the others and leave a crack pattern that repeats around the tool. If the cracks appear at regular angular intervals, look at the spindle before you look at the coolant.
Rigidity is not only the machine. Fixture wear, worn collet nuts and loose vise jaws all add compliance. Check them every few hundred hours, not once a year.
Edge preparation, coolant and wheel conditioning
A sharp as-ground carbide edge is a stress concentrator. Honing it to a 0.02–0.05 mm radius removes the micro-chips left by the wheel and spreads cutting load along the edge. This is standard on cutting tools and often skipped on custom parts that see abrasive service.
Coolant choice follows the operation. Water-based flood coolant handles heat removal in general grinding. Neat oil gives better lubrication and lower friction at the contact point, which matters on fine-grain grades. Molybdenum disulfide or graphite additives in the coolant reduce friction in the contact zone and produce smoother chips, which lowers the mechanical load on the edge.
Wheel conditioning is not optional. Open the bond with a dressing stick, keep the wheel free-cutting, and re-true it when the profile drifts. A glazed wheel is the single fastest way to generate grinding burn.
Do not polish away the geometry you need. Manual polishing helps only when longitudinal and transverse pressure stay balanced. Excess pressure rounds the cutting edge and changes the tool's effective rake.
How to tell a grinding crack from a service crack
Position tells you the origin. Grinding cracks sit parallel to the ground surface, often as a network, and they concentrate on the face that was ground last. Service cracks start at the cutting edge, run inward, and follow the load direction. A crack that appears on a non-functional face is a manufacturing defect, not a wear mechanism.
Timing matters too. Cracks that show up in the first few minutes of cutting come from residual stress or an over-honed edge. Cracks that appear after hours of use come from grade selection, coating adhesion or the cutting parameters.
Fluorescent penetrant inspection finds surface cracks that the eye misses. For critical tooling, a 10× loupe under oblique light is enough to catch the initial signs: a faint line, a lifted chip, or a bright edge that should be dull.
Set a limit. Surface cracks above roughly 0.05 mm deep, or any crack that reaches the cutting edge, mean the tool is scrap. Chasing it with more grinding only moves the crack inward.
Grinding parameters: crack-prone vs. controlled
Ranges for typical WC-Co grades; adjust to your grade and wheel bond
| Variable | Crack-prone | Controlled |
|---|---|---|
| Wheel bond | Hard, glazed, undressed | Softer bond, dressed on schedule |
| Grit size | Coarse 120–150 for finish | Diamond 320–600 mesh for finish |
| Pass depth | 0.05 mm and above | 0.005–0.02 mm finish passes |
| Table speed | Too slow, wheel rubs | 2–5 m/min, matched to wheel |
| Coolant | Mist or intermittent flood | Flood at contact point, 5–8 bar |
| Clamping | Overhang over 3× diameter | Short overhang, rigid support |
| Wheel speed | Above 35 m/s on small wheels | 20–30 m/s for resin bond |
| Spark color | Bright white, heavy shower | Light straw, short sparks |
When to regrind and when to scrap
If the crack is shallow, on a non-cutting face, and the edge has stock left, regrind with reduced pass depth and a freshly dressed wheel. If the crack reaches the cutting edge or runs deeper than 0.05 mm, scrap it. A reground cracked tool fails in the cut, and that is a spindle repair bill, not a tooling bill.
Cracks in cemented carbide tools: common questions
Does a harder carbide grade resist cracking better?
No. Hardness and crack resistance trade against each other. Higher cobalt content, typically 10–12%, gives more toughness and absorbs grinding shock better. Submicron grades with 6% cobalt hold a sharper edge but chip more easily.
Match the grade to the load, not to the spec sheet. If edges fail in service, a tougher grade often outlasts a harder one.
Can a cracked tool be repaired by brazing or welding?
Brazing repairs a joint, not a crack in the carbide body. Heat from the torch adds new residual stress, and the repaired zone will not hold an edge. For solid carbide tooling, regrinding below the crack or replacing the tool is the only reliable path.
How often should a diamond wheel be dressed?
On a fixed schedule tied to material removed, not to the calendar. For most carbide grinding, dressing every 20–40 minutes of contact time keeps the wheel free-cutting. If sparks get brighter or the surface finish degrades, dress sooner.
Do coatings cause or hide cracks?
Coatings do not cause grinding cracks, but they hide them. A cracked substrate under a TiAlN or DLC layer can pass visual inspection. If you suspect grinding damage, inspect before coating, or use penetrant inspection on the finished tool.
Is dry grinding ever acceptable on carbide?
Rarely. Dry grinding works only on very light passes with air blast and a soft wheel, and even then the surface temperature is hard to control. For anything with a cutting edge, flood or neat-oil coolant is the safer choice.
Send us the tool drawing and the failure mode
Tell us the grade, the operation and where the crack appears. We will review the grinding parameters and come back with a quotation and free DFM analysis within 12 hours.
12-hour quote100% inspection±0.005 mmNo minimum order