How to identify carbide.

Read the markings, then feel the weight

Start with the easiest test: look for markings. Reputable manufacturers laser-etch grade codes, part numbers, and logos on insert faces and tool shanks. A 10× loupe turns a faint etch into a readable code — K20, C2, CNMG, a maker's mark — and a code you can read is identification you don't have to guess at. Worn tooling may have lost its etching to years of cutting, which is when you move to the physical tests. Next, feel the weight. Cemented carbide runs about 14 to 15 grams per cubic centimeter against steel's roughly 7.8 — nearly double. Hold the suspect tool in one hand and a known steel tool of similar size in the other. Genuine carbide feels startlingly heavy, like it was machined from something denser than it has any right to be. For a number instead of a feeling, use water displacement: weigh the piece, submerge it, and compute density from the displaced water. Anything above about 12 g/cm³ is mostly carbide; genuine inserts land 13.5 to 15. This is the definitive test — it is also how buyers catch coated-steel counterfeits sold as carbide, which come in around 7.8 to 8.0 like the steel they are.

The magnet test, with nuance

The magnet test is useful but widely misunderstood. Cemented carbide's cobalt binder is magnetic, so most carbide shows some attraction — but the strength of the pull tells a story. A strong pull suggests higher cobalt content, which means lower tungsten content and lower value. A weak pull suggests a lean, tungsten-rich grade. Nickel-binder grades, used where corrosion resistance matters, are non-magnetic and high value. Pure tungsten carbide itself is barely magnetic at all. So the magnet is not a yes/no carbide detector — it is a rough cobalt gauge. A magnet that barely attracts a gray tool is evidence it is unlikely to be steel; a magnet that grabs hard suggests either steel or a high-cobalt grade. Combine the magnet result with the weight test rather than trusting either alone. And remember: stellite and some hardfacing alloys are also weakly magnetic and heavy, so when the tests disagree, the density number wins.

The spark test

The spark test separates carbide from high-speed steel decisively — and it is the test that works even when gold TiN or dark AlTiN coatings hide the underlying color. Touch the tool lightly to a grinding wheel (wear eye protection, and use a wheel you don't mind marking) and watch the spark stream. High-speed steel throws long, energetic, bright orange-to-white streamers that burst and fork at the ends — fireworks, driven by its carbon content. Carbide gives short, slow, dull orange-to-red droplets with almost no bursting — a dying ember rather than fireworks. Stellite falls in between: shorter than HSS, brighter than carbide. The test takes seconds, and once you have seen both patterns side by side you will never confuse them. Two cautions: grind only a small corner of the tool, and never spark-test anything you might want to use again — the test spot is sacrificed.

Sound, scratch, and the counterfeit problem

Two quick confirmation tests round out the sequence. Drop the piece onto a hard surface from a few inches: carbide rings with a high-pitched "tink" while steel answers with a lower "clunk" — the difference in elastic modulus made audible. And carbide scratches glass or a steel file without dulling, while steel cannot return the favor. Visually, real carbide is a dull gray matte; bright, reflective surfaces suggest steel or plating. Now the warning: counterfeit coated-HSS inserts sold as carbide are a real problem in secondary markets. They look right — same shape, same gold coating — but the density test exposes them instantly at ~8 g/cm³ versus ~14. If a deal looks too good, or the "carbide" came from an unknown source, displace water before money changes hands. When every test points the same direction, you know what you have; when they disagree, photograph the markings and the piece and let an experienced buyer review them.

Carbide questions, answered.