Hardness Testing for Machined Parts: Methods, Scales, and Selection
A hardness number is a measurement of resistance to indentation, not a strength figure. This page explains how each common method loads the surface, where it stops being valid, and how to write a callout that a shop can actually meet. Written for design engineers and buyers who sign off on drawings.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
Why hardness testing for machined parts is not a strength test
Hardness testing for machined parts answers one narrow question: how well does this surface resist a loaded indenter? The result is a number tied to a specific method, load, and indenter geometry. It is not tensile strength, and it is not a substitute for a tensile test on the certified material. What it does give you is a fast, cheap check that the heat treat, the alloy, or the cold-work state survived machining.
The relationship between hardness and strength is empirical. For carbon and low-alloy steels, published conversion tables link HRC or HBW to approximate tensile strength, and teams lean on those tables daily. The tables were built for specific material families in specific conditions. Use them as a sanity check on a known steel, not as a design input for an unfamiliar alloy or a cast or additively built part.
Hardness also tracks wear resistance more directly than it tracks toughness. A through-hardened 440C stainless at 58–60 HRC will resist abrasive wear on a sliding surface. The same part at that hardness has almost no ductility left, so a sharp corner under bending load will crack before it deforms. The number tells you which regime you are in. It does not tell you whether the geometry suits it.
One more boundary matters on machined parts. Hardness is measured at a point, or averaged over a small area. A single reading on a carburized gear tooth says nothing about the core, the case depth, or the transition zone. If the drawing only says 58 HRC, a shop can meet it with a case of 0.2 mm and still ship a part that spalls in service.
- 1Indentation resistance, not tensile strengthConversions to strength are empirical and material-family specific.
- 2Wear resistance tracks hardness closelyToughness does not; hardened parts lose ductility as they gain hardness.
- 3A point measurement, not a profileCase depth and core hardness need separate checks.
Rockwell and Rockwell superficial: the default for steel above 2 mm
Rockwell testing per ASTM E18 presses a diamond spheroconical indenter or a 1/16 in (1.588 mm) ball into the surface under a major load of 60, 100, or 150 kgf, then reads the depth of penetration after the load returns to the minor preload. Depth is converted directly to a dial number, so the test takes seconds and needs only a flat, clean surface. That speed is why it dominates incoming inspection and production sampling on machined steel.
The C scale, diamond indenter at 150 kgf, is the standard for hardened steel. The B scale uses the ball at 100 kgf and suits softer, lower-carbon stock and non-ferrous alloys. Scale selection is a frequent error on drawings. A 4140 part at 28 HRC specified on the B scale will read a different number and may confuse the inspection report. Name the scale and the method standard in the callout, not just the number.
Standard Rockwell is valid on parts thicker than roughly 2 mm, or ten times the indentation depth. Below that, the stress field reaches the anvil and the reading runs high or scatters. Rockwell superficial scales, 15N, 30N, and 45N at 15, 30, and 45 kgf, exist for this case. The lighter load keeps the measurement inside thin sections and shallow cases, and the N scale uses the same diamond cone as the C scale.
Superficial scales are the right call for thin washers, small pins, nitrided layers under about 0.5 mm, and light carburizing. If a case is 0.2 mm deep, a standard 150 kgf C-scale test can read the soft core beneath it and return a falsely low number. The part may be perfectly good. The test was wrong.
- 1C scaleDiamond cone, 150 kgf. Hardened steel above about 20 HRC.
- 2B scale1/16 in ball, 100 kgf. Soft steel, brass, aluminium alloys.
- 315N / 30N / 45NSame cone, 15–45 kgf. Thin sections and cases under 0.5 mm.
Brinell: large impressions for castings and coarse structures
Brinell testing per ASTM E10 uses a 10 mm carbide ball under 3000 kgf, or 500 kgf for softer metals, and reports hardness from the diameter of the impression left behind. The impression is wide, so the result averages over a much larger volume of material than Rockwell does. On cast iron, sand castings, forgings, and large sections with coarse or segregated microstructures, that averaging is the whole point. A single Rockwell drop can land on a hard phase and mislead.
The trade-off is damage. A 10 mm impression is a permanent dent roughly 3–4 mm across. It is fine on a raw casting, a billet end, or a non-functional face. It is not acceptable on a finished bearing bore, a sealing surface, or anything that will be polished later. Plan Brinell checks on stock before the finishing operations, or specify a sacrificial pad on the drawing.
Brinell also needs a reasonably flat surface and a part thick enough to hold the impression without the anvil influencing the result. For thin sheet or small parts, the method does not apply. Use superficial Rockwell, Vickers, or a microhardness method instead. Reported values are HBW when a carbide ball is used, and the older HBS designation refers to hardened steel balls that are no longer standard.
For buyers, Brinell is the practical language for castings. A die-cast or sand-cast housing is usually specified in HBW, and the acceptance range is wide because the process is wide. Tightening a Brinell range on a casting to a few points does not improve the casting. It just creates rejects.
- 1Good fitCast irons, sand and die castings, forgings, thick sections.
- 2Poor fitThin sheet, small parts, finished functional surfaces.
- 3Plan the locationTest stock or a sacrificial pad before finishing.
Vickers and microhardness: case depth, welds, and thin features
Vickers testing per ASTM E92 uses a pyramidal diamond and reports HV. Loads run from 1 kgf down to grams, which makes it the tool for anything where the indentation has to be small enough to sit inside a specific feature. Case depth surveys are the classic use. A 0.3 kgf indent is small enough to step through a hardened layer in 0.1 mm increments and plot a hardness profile from surface to core.
That profile is what actually defines a case-hardened part. The effective case depth is usually defined as the distance from the surface to the point where hardness falls to a stated value, often 550 HV or 50 HRC. Without a profile, a single surface reading cannot confirm the case exists at the right depth. Welds get the same treatment. A Vickers traverse across the heat-affected zone shows where softening occurred and how wide it is.
Microhardness at 100–500 gf demands surface preparation that most machine shops do not apply to a production part. A polished surface near Ra 0.1–0.25 μm is normal for this work. On an as-machined face at Ra 1.6 μm, the indenter sits on a surface that is rough relative to the indent, and scatter climbs quickly. Couple that with cold work from turning or grinding, and readings shift by several points.
Vickers is also slow and operator-dependent compared with Rockwell. It belongs on development work, first-article inspection, heat-treat qualification, and failure analysis. Putting a Vickers requirement on every part in a 10,000-piece run adds cost without adding control. Specify the method where the information lives.
- 1Case depthStep in 0.1 mm increments, plot HV against distance.
- 2Weld HAZTraverse across the zone to find softening width.
- 3Prep requirementNear Ra 0.1–0.25 μm for loads under 500 gf.
Leeb rebound and portable testers: screening, not acceptance
Leeb rebound testers per ASTM A956 drop a small impact body onto the surface and measure rebound velocity. The device is handheld, the test is quick, and it reaches large weldments and installed parts that will never fit on a bench tester. That convenience is why portable units are common on site. It is also why their results are often over-trusted.
Leeb is a comparative method. The reading depends on part mass and rigidity. A 2 kg part needs rigid support or a coupling medium to read consistently, and thin walls flex under the impact and read low. Surface finish, curvature, and the direction of the impact body all shift the number. Manufacturer tables convert Leeb to HRC, but those tables assume conditions that field work rarely matches.
The honest role for Leeb is screening and sorting. It separates a mixed bin of heat-treated and untreated parts in minutes. It flags a suspect area on a large weldment that then gets a bench check. It is weak as the sole acceptance record for a drawing callout. If a purchase order says 32–36 HRC, a bench Rockwell at a qualified lab is what closes the argument.
Ultrasonic contact impedance testers follow similar logic. They are useful on large or assembled parts and on some coatings, and they carry the same caveat: the correlation to bench hardness depends on material, geometry, and surface. Treat any portable reading as a relative indicator unless it has been correlated to a bench method on the same material and geometry.
- 1Use forSorting, field checks, large weldments, installed parts.
- 2Do not use asThe sole acceptance record against a drawing callout.
- 3Correlate firstSame material and geometry as the bench reference.
Surface preparation and where the indent goes
Every method assumes a surface flatter and cleaner than a typical as-machined face. An Ra 1.6 μm machined surface is usually adequate for Rockwell, but scale, decarburization, and cold work from turning or grinding can shift readings by several points. The standard fix is to grind or lightly polish the test area to about Ra 0.8 μm, then take at least three readings and average them. On a hardened surface, remove enough material to get under the decarburized layer.
Placement rules exist for a reason. Keep indents at least 2.5 times the impression diameter from any edge, and keep part thickness at least ten times the indentation depth. On a 3 mm hardened washer, a 150 kgf C-scale test can be invalid because the stress field reaches the anvil and the reading runs high. The part is fine. The test setup is not.
Spacing between indents matters too. Readings taken closer than about three impression diameters interact through the strained zone around each indent, and the second reading is biased. On small parts, this forces a choice between fewer valid indents and a smaller scale. Take the smaller scale.
Finally, decide where on the part the hardness actually needs to be. A shaft may be induction hardened on the bearing journals and left soft in the middle for toughness. A single callout on the drawing does not express that. Mark the zones, give a value for each, and state the method.
- 1Edge distanceAt least 2.5 × impression diameter from any edge.
- 2ThicknessAt least 10 × indentation depth under the indenter.
- 3SpacingAt least 3 × impression diameter between readings.
Scale conversion and its limits
Conversion tables between HRC, HBW, HV, and tensile strength are convenient and frequently misused. The published tables, including ASTM E140, were derived from specific material groups under specific conditions. They are reasonably reliable within a family such as quenched and tempered carbon steel. They become unreliable across families, at very high hardness, and on cast, welded, or additively built material.
The practical rule is to test in the scale you specified. If the drawing says 58–62 HRC, measure in HRC. Do not measure in HV, convert, and accept or reject on the converted number. Conversion belongs in reporting and cross-reference, not in the acceptance decision. Where a conversion is unavoidable, state the conversion standard on the report so the reader knows what was done.
Hardness values below roughly 20 HRC convert poorly in both directions because the scales compress at the soft end. Above about 60 HRC, the diamond indenter starts to damage and readings drift. Neither end is a good place to build a tight tolerance. If the design is that sensitive, the material or the process is probably the wrong choice.
One more limit worth stating plainly. Hardness cannot detect a wrong alloy that happens to hit the same number. A 1018 part and a 4140 part can both be made to read 30 HRC by different heat treatments, and the hardness test will pass both. Alloy verification needs a different check, such as PMI or a mill certificate review.
- 1Test in the specified scaleConvert for reporting only, not for accept or reject.
- 2Weak at the endsBelow 20 HRC and above 60 HRC, conversions drift.
- 3Not an alloy checkA wrong alloy can hit the same hardness number.
Hardness methods compared by load, use, and damage
Loads and applications are per ASTM E18, E10, E92, and A956.
| Method | Typical load | Best for | Main limit |
|---|---|---|---|
| Rockwell C | 150 kgf | Hardened steel above 20 HRC | Needs thickness over about 2 mm |
| Rockwell B | 100 kgf | Soft steel, brass, aluminium | Not for hardened parts |
| Rockwell superficial | 15–45 kgf | Thin parts, cases under 0.5 mm | Shallow indent, surface-sensitive |
| Brinell HBW | 500–3000 kgf | Castings, forgings, coarse structures | Large visible impression |
| Vickers HV | 1 kgf to grams | Case depth, weld HAZ, small features | Slow, needs polished surface |
| Leeb rebound | Portable impact | Field screening, large weldments | Comparative, mass-dependent |
The selection rule
For hardened steel above 2 mm thick, specify Rockwell C and stop there. For thin parts or cases under 0.5 mm, specify Rockwell superficial 15N or 30N. For castings and coarse sections, use Brinell HBW and test on stock. Use Vickers only where you need a case depth profile or a weld traverse, and treat Leeb as screening, never as the acceptance record.
Questions engineers ask about hardness on drawings
Can a shop meet a hardness callout without a heat-treat certificate?
A hardness reading confirms the surface condition at the point tested. It does not confirm which alloy was used, how the part was heat treated, or whether the treatment was uniform through the section.
For anything safety-related or lot-controlled, pair the hardness check with a mill certificate and, where the alloy matters, a PMI check. The hardness test is a fast gate, not a full material record.
Why did my part pass at the shop and fail at incoming inspection?
The usual causes are different scales, different surface preparation, or a different test location. A reading on a decarburized surface, a ground surface, or an edge region can differ by several points from a reading on a prepared flat face.
Compare the method, the load, the scale, and the location before disputing the result. A single reading is also weak evidence; three readings averaged is the normal basis.
Is hardness testing required on every part in a production run?
Rarely. Hardness testing is usually destructive or at least surface-damaging, so it is applied on samples. The sampling plan should match the risk: first article plus periodic checks for a stable heat-treat process, more frequent checks when a new lot of material or a new heat-treat batch enters.
For a 10,000-piece run, testing every part adds cost and damage without adding much control. Test the process, not each piece, and keep the records.
What does 40 HRC mean in terms of tensile strength?
For quenched and tempered carbon and low-alloy steel, published tables put 40 HRC in the range of roughly 1,250–1,300 MPa tensile. Treat that as an estimate.
The conversion assumes a specific steel family and condition. It is a reasonable cross-check on a known material and a poor basis for sizing a part in an unfamiliar alloy.
Can hardness testing be done on anodized or plated parts?
Not meaningfully on the coating itself with bench methods. An anodized layer or a plated layer is thin relative to a standard Rockwell or Brinell indent, so the indenter reads through to the substrate.
Test the substrate before finishing. If coating hardness matters, that is a separate microhardness measurement with its own preparation requirements and its own specification.
How do we specify hardness on a drawing for a part with hard and soft zones?
Mark each zone on the drawing and give a value, a scale, and a method standard for each. State the depth of the hardened zone if it matters, and define the effective case depth criterion, such as 550 HV.
A single global callout on a selectively hardened part will be met somewhere and missed elsewhere. Zoning the drawing removes the ambiguity.
Send us the drawing and the hardness callout
We review the callout against the material and geometry, flag scales that cannot be met on the part as drawn, and quote with the inspection method stated.
12-hour quote100% inspectionDFM feedback included