Rockwell vs Brinell vs Vickers is not a contest between three ways of measuring the same thing. The three tests press different indenters into the metal under different forces and read different features of the dent, and the standard that publishes the conversion tables between them — ASTM E140 — states in the title of every table that its values are approximate. Convert a number, print it on a spec sheet, and you have quietly turned a measurement into an estimate.
That is where most hardness disputes on export orders start. Not with a lab being wrong. With two labs being right about different things, and a spec sheet that only had room for one number.
Rockwell vs Brinell vs Vickers: what each test actually does
Three self-contained definitions, because the difference is in the last few words of each one.
Rockwell hardness is the depth of the permanent indentation left by a defined indenter under a defined force, read directly off the machine as a dimensionless number on a lettered scale (HRC, HRB and others).
Brinell hardness is the test force divided by the curved surface area of the indentation left by a hard metal ball, calculated from the diameter of the dent measured under a microscope.
Vickers hardness is the test force divided by the surface area of the indentation left by a 136-degree square-based diamond pyramid, calculated from the mean of the dent's two diagonals.
| Rockwell C (HRC) | Brinell (HBW) | Vickers (HV) | |
|---|---|---|---|
| Governing standards | ASTM E18 / ISO 6508 | ASTM E10 / ISO 6506 | ASTM E92 and E384 / ISO 6507 |
| Indenter | 120-degree diamond cone | hard metal ball, D = 10, 5, 2.5 or 1 mm | 136-degree square-based diamond pyramid |
| Preliminary force | 10 kgf | not applicable | not applicable |
| Typical total force | 150 kgf | 29.42 kN with the 10 mm ball for steel (load factor 30) | preferred macro values 49, 98, 196, 294, 490, 980 N |
| Measured feature | indentation depth | indentation diameter, optically | mean of the two diagonals, optically |
| Result appears | immediately, on the machine | after the dent is measured | after the dent is measured |
| Practical ceiling | none in normal steel work | about 653 HBW for steel and iron | none in practice |
| Minimum specimen thickness | per ASTM E18, based on indent depth | at least 8x the indentation depth | at least 1.5x the indentation diagonal |
| Surface preparation | light | ground | ground and polished |
| Strongest at | hardened steel, fast production QC | castings, forgings, coarse or non-uniform structure | thin sections, coatings, case depth, small parts |
For softer material the Rockwell family switches indenter rather than method: HRBW uses a 1/16-inch ball at 100 kgf and is where unhardened steels and copper alloys are read.
Why there is no exact conversion between the three
ASTM E140 exists because everyone needs to convert. It also tells you not to trust the result too far. The tests do not measure the same combination of material properties, so conversion between them is an approximation, and each table in the standard is valid only for the material family it names — non-austenitic steels, austenitic stainless sheet, nickel alloys, cartridge brass and so on. Take a table built for carbon steel and apply it to an austenitic stainless casting and the error is not a rounding error.
Two anchor points from published E140-based charts, so the shape of the relationship is concrete: 300 HBW is about 30.5 HRC, and 500 HBW is about 49.1 HRC.
Here is what that means on a purchase order. Your drawing says 45 HRC. Your mill certificate reports 421 HBW because that is the machine the foundry owns. You have not met the specification — you have offered a conversion of it. When the buyer's incoming inspection runs Rockwell C and reads 43, there is no way to win the argument, because the disagreement was manufactured by the conversion rather than by the metal.
A hardness value is a measurement only in the scale it was measured in. In any other scale it is an estimate. Specify and report in the scale the part will actually be inspected in, and the entire class of dispute disappears.
Which test your part can even take
In practice, Rockwell vs Brinell vs Vickers is decided by the part, not by preference.
Thickness. Brinell needs the specimen to be at least eight times thicker than the indentation depth; Vickers needs at least 1.5 times the indentation diagonal. A 1.5 mm stamped bracket cannot take a 10 mm Brinell ball at 3,000 kgf — the dent goes through and marks the far face, and the reading is worthless.
Hardness ceiling. Brinell stops being usable at roughly 653 HBW for steel and iron. Push past it and the ball itself deforms and material piles up at the rim of the dent, so the diameter you measure is not the diameter the calculation assumes.
Structure. A 10 mm Brinell ball leaves an indent millimetres across and averages over many grains. That is exactly why castings and forgings are specified in HBW. Put a micro-Vickers indent into the same casting and it can land inside a single grain or on a graphite flake and return a number that is real, repeatable and useless as a lot average.
Case-hardened parts. The surface is hard, the core is not. A Brinell ball breaks through the case and reads a blend. Case depth is a Vickers job: section the part, polish it, and run a hardness traverse from the surface inward.
Throughput. A Brinell cycle runs 30 to 60 seconds before anyone has measured anything, and then the dent has to be read optically. Rockwell puts a number on the display. On a line checking every twentieth part, that difference decides the method.
Where the number goes wrong on an export order
Six failure modes, in rough order of how often they turn into a claim.
- Scale mismatch on paper. The drawing is in HRC, the certificate is in HBW, and nobody converted with the right table — or converted at all.
- Wrong method for a case-hardened part. The part is correct; the test read through the case and reported a failure.
- A number with no scale letter. "Hardness 55" is not a specification. It is the same failure as writing "Shore 70" on a rubber part, covered in Shore A vs Shore D hardness.
- A single value with no tolerance. Hardness is a distribution across a lot and across a part face. A spec without a range invites a rejection on the first outlier.
- No test location. Surface, mid-radius and core can differ by tens of points on the same component. If the drawing does not say where, the two labs will not test the same place.
- "HB" with no ball material. Modern Brinell testing uses a tungsten carbide ball and is written HBW; the old steel-ball notation is retired. "HB 300" leaves the reader guessing which era the number came from.
How to write a hardness callout that survives inspection
| Written as | What goes wrong | Write instead |
|---|---|---|
| Hardness 55 | no scale, so no meaning | 55 +/- 2 HRC |
| 58-62 HRC on a 1 mm sheet | violates the minimum thickness rule for the method | 650-750 HV1, measured on the surface |
| HB 300 | obsolete notation, ball material unstated | 300 HBW 10/3000 |
| Hardened and tempered | a process, not a measurement | method, scale, value, tolerance and test location |
| 45 HRC (converted from Brinell) | an estimate presented as a measurement | report the scale you measured in; add the conversion as a note only |
| Hardness per drawing | pushes the ambiguity downstream | repeat the full callout on the certificate |
A complete callout has five parts: method and scale, nominal value, tolerance, test location, and the standard it is run to. "550 +/- 25 HV10 on the ground face, per ISO 6507-1" cannot be misread. "Hardness 550" can be misread six ways.
The same discipline applies to the neighbouring numbers on a metal spec sheet — surface finish is a twin case, where the same figure means two different surfaces depending on the parameter, as in Ra vs Rz surface roughness. The broader checklist for the rest of the sheet is in how to label metal product specifications.
Next steps
Three moves, in the order that removes the most risk per hour spent.
- Ask the buyer which scale their incoming inspection uses, before the first shipment. One email removes the single largest cause of hardness claims. Then align your drawing and your certificate to that scale rather than converting at the border.
- Audit your existing drawings for incomplete callouts. Every "hardness 55" and every "HB 300" is a dispute waiting for a slow month. Fix the notation once; it costs nothing and it never comes back.
- Put the number where the buyer reads it, not only where the file lives. On a catalogue page, a marketplace listing or an RFQ reply, buyers do not open the drawing PDF; they look at the image. A spec sheet built as an annotated image carries the hardness callout next to the measured dimensions that were taken from the part's real geometry — snapped to the actual edges of the object in frame rather than typed on from memory — and exported at the pixel size each channel requires so the small text survives. That is the difference between a spec a buyer can act on and a spec they have to ask you about. It is also why a generated illustration is the wrong tool here: it will letter a plausible number onto a part it never measured.
FAQ
Which is better, Rockwell or Brinell?
Neither, in the abstract — Rockwell vs Brinell vs Vickers is a fit question, not a ranking. Rockwell is faster and needs less surface preparation, which makes it the production-line choice for hardened steel. Brinell's large indent averages over a coarse or non-uniform structure, which is why castings and forgings are specified in HBW. Pick the one that matches the part's structure and thickness, then specify in that scale.
Can you convert HRC to HB accurately?
No, not in the sense that "accurately" usually implies. ASTM E140 publishes the conversion tables and states that the values are approximate, because the tests measure different combinations of material properties. The tables are also material-specific: the one for non-austenitic steels does not apply to austenitic stainless or to nickel alloys.
What hardness test is used for thin parts?
Vickers, in most cases. The specimen only needs to be about 1.5 times the indentation diagonal in thickness, and the force can be scaled down to grams for thin sheet, coatings and case-depth traverses. Brinell requires eight times the indentation depth, which rules it out on thin sections.
Why does my mill certificate show a different hardness than the buyer's lab?
Usually one of three things: the two labs used different scales and someone converted between them, the part is case-hardened and one test read through the case, or the two tests were run at different locations on the part. Specifying the test location and the scale on the drawing eliminates all three.
How do you write hardness on a technical spec sheet?
Method and scale, nominal value, tolerance, test location, and the standard — for example "550 +/- 25 HV10 on the ground face, per ISO 6507-1", or "300 HBW 10/3000". Never a bare number, and never a converted value presented as the measured one.
Sources & References
- ASTM E18 — Standard Test Methods for Rockwell Hardness of Metallic Materials
- ASTM E10 — Standard Test Method for Brinell Hardness of Metallic Materials
- ASTM E92 — Standard Test Methods for Vickers Hardness and Knoop Hardness of Metallic Materials
- ASTM E140 — Standard Hardness Conversion Tables for Metals: the tables are titled as approximate relationships and are material-specific
- ZwickRoell — Rockwell hardness test to ISO 6508: indenters, preliminary and total test forces by scale
- ZwickRoell — Brinell hardness test to ISO 6506: ball diameters, load factors, the 653 HBW ceiling and the 8x thickness rule
- ZwickRoell — Vickers hardness test to ISO 6507: 136-degree indenter, preferred test forces and surface preparation
- Metkon — Hardness conversion tables: 300 HBW is about 30.5 HRC and 500 HBW is about 49.1 HRC
- NIST — Hardness standardization programme: national standards and reference materials for hardness scales